Freedom of Speech and Appellate and Summary Judgment Review in Copyright Cases

Prof. Eugene Volokh, UCLA Law School *
Brett McDonnell **

(107 Yale L.J. 2431 (1998))


I.      Why Independent Judgment Review Is Mandated
      A.      Freedom of Speech and Copyright
      B.      Freedom of Speech and Appellate Review
      C.      Appellate Review and Copyright
      D.      Appellate Review in Cases Won By Defendants
      E.      Applying These Principles to Summary Judgment and Motions for J.N.O.V.
      F.      Lawyers Should at Least Ask for Independent Review
II.      Is Copyright Different?
      A.      The First Amendment Interest
            1.      Property Rights
            2.      Private Enforcement
            3.      Content-Neutrality
            4.      Subject Matter of the Jeopardized Speech
            5.      It's All Costs of Doing Business
            6.      Copyright Law Furthers Free Speech Values
            7.      The Copyright Clause
            8.      Importance of the Government Interest
      B.      Is Trying to Refine the Idea/Expression Dichotomy Pointless?
            1.      Lower Courts' Use of Prior Cases as Benchmarks
            2.      Lawyers' Use of Prior Cases as Benchmarks
            3.      The Supreme Court's Views in Other Areas
            4.      Independent Appellate Review in Those Circuits That Already Do It
      C.      Against Special Pleading for Copyright
III.      Reconsidering Bose?
IV.  First Amendment Due Prcoess and Copyright Law


              Copyright law restricts speech.  It restricts what writers may write, what painters may paint, what composers may create.  It doesn't limit itself to slavish copying, but prohibits people from creating entirely new works, so long as those works use -- even if only in part -- another's expression. 1

              Of course, the Supreme Court has held that copyright law is a valid speech restriction.  Because the law stimulates entry into the marketplace of ideas, and because the law prohibits only the use of others' expression, not their ideas or the facts they've uncovered, the Copyright Act does not violate the First Amendment. 2

              Nonetheless, as the Court has time and again held, certain procedural safeguards must accompany even substantively valid speech restrictions.  One such safeguard is independent judicial review, by appellate courts after trial, by the trial court after trial, and by the trial court on summary judgment.  Under Bose Corp. v. Consumers Union, 3 appellate courts can't just turn over vague phrases such as "actual malice" or "incitement" or "expression, as opposed to idea" to factfinders, and then defer to the factfinders' conclusions about what constitutes libel or incitement or copyright infringement.  Rather, courts must "conduct[] an independent review of the record both [(1)] to be sure that the speech in question actually falls within the unprotected category and [(2)] to confine the perimeters of any unprotected category within acceptably narrow limits in an effort to ensure that protected expression will not be inhibited." 4  Lower courts have properly accepted this principle for trial court review on motions for summary judgment and for judgment notwithstanding the verdict. 5

              In theory -- a theory accepted by the Supreme Court as a principle of constitutional law -- such independent review prevents prejudiced or erroneous deprivation of constitutional rights by factfinders.  If a factfinder erroneously concludes that your book infringes someone else's book, the factfinder hasn't just made a legal mistake:  It has made a mistake of constitutional magnitude -- it has deprived you of your First Amendment right to write your own expression, even when based on another's idea.  Courts must, Bose holds, police factfinders' decisions to protect against such mistakes.

              Beyond this, independent review is also supposed to help prevent future mistakes, by making the lines in free speech law clearer and more administrable.  Judicial review is part of the "evolutionary process of common law adjudication" which "gives content" to legal rules. 6  As courts see more cases of a particular type, they can refine the line between protected speech (such as nonobscene art, innocent error, or copying of ideas) and unprotected speech (such as obscenity, punishable libel, or copying of expression).  They might create new subrules that clarify the meaning of the rules, both for the benefit of future courts and of future speakers.  Or they might provide benchmarks, against which future courts can compare and contrast new fact patterns.

              In Part I, we explain why Bose Corp. v. Consumers Union compels independent review of "substantial similarity of expression" determinations. 7  Though the great majority of circuits have held, without considering free speech issues, that such determinations should be reviewed only for clear error, we believe these circuits are mistaken.  The doctrinal demands of Bose are quite clear.  In Part II, we argue that nothing special about copyright cases justifies departing from the independent judgment rule.  In light of this, giving copyright law a free ride not given other speech restrictions is wrong, and corrosive of people's respect for free speech generally.

              In Part III, we ask whether this result -- and Bose itself -- makes sense.  The Supreme Court's "First Amendment Due Process" 8 jurisprudence has been a pragmatic, largely seat-of-the-pants, judgment about the real-world impact of various sorts of procedural rules, be they independent review, punitive damages, or what have you.  Could Bose be mistaken, either as applied to copyright law or generally?  Has the Court gone too far in constitutionalizating procedure as well as substance in free speech cases?  Should it return to treating speech-based claims the same way that other claims -- negligence claims, contract claims, and the like -- are treated?

              We believe that the Court's judgments are probably correct, and that in any event, the Court's holding in Bose at least switches the burden of proof to those who would argue against it.  In the absence of a persuasive argument for evading independent review in copyright cases, courts deciding such cases must follow Bose.  The error-correcting and law-clarifying benefits of independent review exceed the costs imposed on the system and on litigants by the increased likelihood of appeals. 9  Nonetheless, it's important to think skeptically about such broad but unproven judicial pronouncements regarding the likely effects of law (here rules of judicial review) on human action (here decisions by lower courts and by creators).  If people believe that independent review is inappropriate in copyright cases, this might be an opportunity to rethink Bose generally.

              Finally, in Part IV, we suggest that lawyers and scholars should seriously consider whether other "First Amendment Due Process" rules -- which require, for instance, proof by clear and convincing evidence or a prohibition on punitive damages -- likewise apply to copyright cases.  This issue deserves more attention than it has so far gotten.

I.Why Independent Judgment Review Is Mandated

A.Freedom of Speech and Copyright

              Copyright law restricts speech:  It restricts you from writing, singing, painting, or otherwise communicating what you please. 10  If your speech copies ours, and if the copying uses our "expression," not merely our ideas or facts, it can be enjoined and punished, civilly and sometimes criminally. 11

              And copyright law applies to creative adaptation, not just to literal copying.  Rap musicians are restricted from including "samples" of others' music in their own songs. 12  Artists are forbidden from creating artworks that are too similar to others' art. 13  Writers are barred from creating works -- even works based on real events -- whose plots are too similar to what others have done. 14  Copyright law is a serious restriction on speakers' desire to express themselves the way they want.

              Harper & Row v. Nation Enterprises made clear that speech that infringes another's copyright isn't constitutionally shielded from copyright law.  Copyright is itself an "engine of free expression," because it "supplies the economic incentive to create and disseminate ideas"; and this justifies copyright law's limitation on use of others' expression. 15

              But the Court strongly implied that this rationale wouldn't justify restrictions on copying facts or ideas.  The Court characterized "the Copyright Act's distinction between copyrightable expression and uncopyrightable facts and ideas" as a "First Amendment protection[]."  It pointed out that "[n]o author may copyright his ideas or the facts he narrates," and cited Justice Brennan's opinion in New York Times Co. v. United States, 16 which said that "[c]opyright laws are not restrictions on freedom of speech as copyright protects only form of expression and not the ideas expressed."  And it stressed that it would be an "abuse of the copyright owner's monopoly" for copyright law to become "an instrument to suppress facts."

              Harper & Row thus suggests that the line between using others' expression and using their ideas is of First Amendment significance.  Speech communicating facts and ideas using expression that's "substantially similar" to someone else's expression is constitutionally unprotected.  Speech communicating the same facts and ideas in other ways is constitutionally protected. 17

              Such a dividing line makes good constitutional sense.  The free speech principle may tolerate certain limits on how someone expresses an idea or a fact, but -- whether one sees the principle as primarily concerned with protecting self-expression or with fostering democratic discourse or the marketplace of ideas -- it cannot tolerate restrictions on communicating ideas and facts as such.  When you express an idea someone else pioneered or discuss facts that others have uncovered, you might be free-riding on their hard work, but it's a free-riding we must allow. 18

              Thus, we'll assume what Harper & Row strongly suggested, and what lower courts have reaffirmed:  Speech that copies another's expression can be restricted, but speech that borrows facts or ideas is constitutionally protected.

B.Freedom of Speech and Appellate Review

              Speech copying another's expression isn't, of course, the only category of speech unprotected by the Free Speech Clause.  Fighting words, obscenity, and libel, for example, are also generally unprotected.  For each category, the Court has set forth rules defining the category's boundaries:  For instance, defamatory statements about public figures are actionable only if made with "actual malice" -- knowledge of or reckless disregard of their falsity. 19

              But these rules are not self-explanatory, and it's not enough for appellate courts just to announce the rules and leave them to judges and juries to apply.  "Providing triers of fact with a general description of the type of communication whose content is unworthy of protection has not, in and of itself, served sufficiently to narrow the category, nor served to eliminate the danger that decisions by triers of fact may inhibit the expression of protected ideas." 20

              Therefore, the Court has held, courts must independently review judgments that a certain statement is unprotected.  In part, this simply prevents unconstitutional results:  Because erroneous denial of constitutional protection is a violation of constitutional rights, courts must "exercise [independent] review in order to preserve the precious liberties established and ordained by the Constitution." 21

              But beyond that, independent review is also supposed to make the rule clearer for future cases.  Independent review should help "confine the perimeters of any unprotected category within acceptably narrow limits in an effort to ensure that protected expression will not be inhibited." 22  The content of many Free Speech Clause rules "is not revealed simply by [the rule's] literal text"; 23 rather, the rules must be "given meaning through the evolutionary process of common law adjudication." 24  Therefore, appellate judges, "as expositors of the Constitution, must independently decide whether the evidence in the record is sufficient to cross the constitutional threshold." 25

              If appellate courts review decisions only for clear error, not independently, the "evolutionary process of common law adjudication" is substantially stunted.  Instead of marking out two areas -- copying of idea and copying of expression -- clear error review marks out three areas:

              (1) copying that any reasonable factfinder would conclude only copies ideas,

              (2) copying that any reasonable factfinder would conclude only copies expression, and

              (3) copying on which reasonable factfinders could disagree. The third area is very big, and decisions that fall within this area give little guidance.  A factfinder doesn't get much value from a precedent saying "Reasonable factfinders could disagree whether speech X is substantially similar in expression to speech Y." 26  Likewise, such a precedent gives little guidance to speakers who want to know what they can say and what they can't -- and, as the Court has recognized, when the rules are so vague, many speakers will "steer far wider of the forbidden zone." 27

              The above quotes praising independent review come from Bose Corp. v. Consumers Union, a libel case, but the reasoning applies equally to copyright law: * In both cases, some speech is protected and some isn't. * In both cases, the factfinder may misclassify the speech as unprotected -- erroneously conclude that it was said with actual malice, or erroneously conclude that it used another's expression. * In both cases, the rule's literal text provides little guidance without case-by-case elaboration. 28

              Bose made clear that its rule applies generally, beyond libel, to judgments that a certain kind of speech is unprotected. 29  Post-Bose cases have faithfully applied Bose to alleged obscenity, 30 incitement, 31 negligent publication of criminal solicitation, 32 speech by lawyers supposedly interfering with the administration of justice, 33 government employee speech, 34 speech in a possibly nonpublic forum, 35 commercial speech, 36 and content-neutral speech restrictions. 37

              If anything, independent judgment review seems particularly proper in copyright cases.  "Substantial similarity of expression" is an amorphous term.  It's at least as vague as "prurient interest" and "patently offensive" (elements of the obscenity test), and probably vaguer than "reckless disregard" (part of the libel test) and "likely to provoke the average person to retaliation" (part of the fighting words test). 38  No longstanding social consensus tells us what's "idea" and what's "expression"; no intuitively obvious line divides the two categories.  Under Bose, this is precisely the sort of test courts must police and clarify through case-by-case adjudication, not leave entirely to the ad hoc decisions of judges and jurors. 39

C.Appellate Review and Copyright

              Today, many circuit courts review substantial-similarity-of-expression findings made in bench trials for clear error, 40 though the Second Circuit, perhaps joined by the Eleventh, applies independent judgment (also known in this context as de novo review). 41  All circuits review copyright jury verdicts by asking whether any reasonable jury could have reached the verdict. 42

              In fact, the de novo standard of review may often be proper even as a nonconstitutional matter.  Most courts have treated substantial similarity of expression as a factual question, which is the main reason they review it for clear error. 43  The substantial-similarity-of-expression judgment, though, is really an application of law to fact, also known as a mixed question of law and fact.  When one compares, say, two books, the facts of their contents are uncontroverted; the issue is whether the expression in those contents passes the legal test of substantial similarity.  Most circuits hold that mixed questions may be reviewed de novo, and that Rule 52(a) dictates clear error review only for purely factual questions. 44  General jury verdicts, however, are always reviewed deferentially unless the Constitution commands otherwise. 45

              How would a de novo standard work in practice?  Say a factfinder finds that a defendant's work is substantially similar in its expression to the plaintiff's work.  And say the court of appeals, applying independent judgment, disagrees. 46

              This decision will then become a benchmark against which future courts -- and, better yet, future creators and publishers -- can compare and contrast their cases.  Of course, no two fact patterns are identical.  But the data points may add up.  As the Supreme Court said when adopting an independent review standard for similarly vague Fourth Amendment probable cause determinations, "even where one case may not squarely control another one, the two decisions when viewed together may usefully add to the body of law on the subject." 47  With each new binding decision the rule becomes a little clearer.

              Independent judgment review needn't -- and can't -- reexamine all the factual findings involved in the lower court's decision.  The appellate court may, for instance, defer to the factfinder's judgments about witness credibility. 48  In jury cases, the court will generally have to assume that the jurors believed the winning side's factual claims.

              But even if this is done, the question will remain:  Accepting the winner's story about the historical facts, are the two works substantially similar in their expression?  The appellate court can make this decision at least as well as a jury or a trial judge. 49

              Some skeptics suggest that in practice the standard of review matters little -- that judges will manipulate the standard to reach the result they want.  We disagree.  Doubtless such manipulation sometimes happens, but in our experience courts generally do take the standard of review seriously.  Courts certainly say that standards of review matter, and it seems standards of review must sometimes make a difference. 50

              Without independent judgment review, cases give little guidance about what's allowed and what's forbidden. 51  Precedents are set only when an appellate court concludes that no reasonable factfinder could find liability or fail to find liability.  By definition, this happens only in rather extreme cases -- cases where a court could find that a jury would have to be, in one judge's words, "drunk or crazy" to conclude that the defendant's speech was substantially similar in its expression to the plaintiff's. 52  That's not an easy standard to meet.

D.Appellate Review in Cases Won By Defendants

              Bose did leave a significant question unresolved:  Is independent judgment review proper if the defendant wins at trial?  The lower courts are split on this.

              Some courts stress that independent judgment review aims at developing and refining the constitutional rules.  This development would happen whoever won below, suggesting independent judgment review should apply symmetrically. 53

              Other courts stress the other aspect of the Bose holding -- appellate review decreases the chances that constitutionally protected speech would be erroneously punished.  Under this view, when the free speech claimant wins below, there's no risk that the factfinder has erroneously abridged a constitutional right. 54  Indeed, independent appellate review in this situation increases the chance of erroneously punishing protected speech, though it decreases the chance of erroneously protecting unprotected speech.  Moreover, the argument goes, courts can't adopt independent judgment review just for prudential reasons.  Under Federal Rule of Civil Procedure 52(a), appellate courts must review factual findings for clear error, unless the Constitution commands otherwise. 55  The same holds true for jury trials, under the Seventh Amendment. 56

              In our view, independent judgment review of the idea-expression decision is valuable even when the defendant won at trial:  Whoever won, appellate independent judgment should produce more refinement of the legal standard, something that Bose says is constitutionally valuable.  Moreover, a symmetric rule is fairer to plaintiffs.  Copyright plaintiffs' claims aren't claims of constitutional right, but they're certainly important; as Harper & Row pointed out, copyright law itself serves First Amendment goals.

              For review of bench trials, the symmetric approach can be used whether or not one concludes that Bose requires it as a constitutional matter; as we mention above, most circuits hold that decisions involving application of law to fact may be reviewed de novo without running afoul of Rule 52(a). 57  And the rationale for reviewing mixed questions de novo -- that questions that involve "strik[ing] a balance between two sometimes conflicting societal values" and that are therefore "of clear precedential importance" 58 should be decided by appellate courts -- applies well in the copyright context.

              Independent judgment review probably can't work, regardless of how one reads Bose, when the court is reviewing a jury's general verdict for a defendant.  Copyright claims involve subsidiary factual inquiries.  When the defendant wins, it might be either because the jury concluded that there was no substantial similarity of expression, or (for instance) because the jury concluded that there was independent creation and thus no copying at all.  The court of appeals has no way of knowing the real reason, and thus can't reverse even if it thinks there was substantial similarity of expression.

              Review of special verdicts, however, is possible.  In particular, the Seventh Amendment, which bars only reexamination of "fact[s] tried by a jury," 59 probably doesn't prevent appellate courts from reviewing mixed questions of law and fact. 60

E.Applying These Principles to Summary Judgment and Motions for J.N.O.V.

              The same principles apply when trial courts review motions for summary judgment and for judgment notwithstanding the verdict:  The court must independently decide whether the two works are indeed substantially similar in their expression, and not just whether a reasonable jury could so find.

              This makes sense doctrinally, practically, and theoretically.  Doctrinally, federal courts have held that Bose applies to decisions on summary judgment and motion for j.n.o.v. 61  Practically, for j.n.o.v. motions, the court of appeals would in any event -- if an appeal is filed -- review the matter independently.  There's no reason to deny the losing party this independent review until the appellate court rules; and providing this review up front might better reach the right result even when no appeal is filed. 62

              The same goes for summary judgment:  If, resolving all the underlying factual claims in the movant's favor, the court still concludes that there's no sufficient similarity of expression, there's no reason to delay this judgment until the j.n.o.v. motion or until review by the court of appeals.  In fact, having trial courts review the evidence deferentially and then having appellate courts review it independently would just lead to needless reversals, reversals that might have been avoided if the trial courts reviewed the evidence independently to begin with.

              Theoretically, at least one of the underlying principles of Bose applies fully to trial court review:  Trial courts as well as courts of appeals have a duty to prevent erroneous denials of constitutional protection.  The other principle -- the notion that judicial decisions can help clarify the law -- is somewhat less applicable, but still retains considerable force.  In copyright practice, district court decisions are often quite influential: 63  Because copyright lawsuits are mainly about money, losing parties often settle or just give up rather than appeal.  Even district court decisions can thus be influential benchmarks that help "give[] meaning [to the rules] through the evolutionary process of common law adjudication." 64

              In any event, as the cases cited above show, the principle that Bose generally applies on summary judgment and motion for judgment j.n.o.v. is well-entrenched in the caselaw.  The burden is on those who would carve out a special copyright exception; and for the reasons we discuss below, we are skeptical that this burden can be met.

F.Lawyers Should at Least Ask for Independent Review

              Finally, whether or not the above claims are sure winners, they are at least colorable enough that lawyers who lose at trial ought to raise them.  Stare decisis doesn't prevent courts from adopting this approach, even if they have in the past reviewed substantial-similarity-of-expression findings only for clear error; none of the cases adopting a clear error standard considered Bose.  When a new argument is raised that wasn't considered in a prior case, a court isn't bound by the prior decision. 65  And our proposal is hardly radical:  As we discuss below in Part II.B.4, some circuits have indeed accepted independent review with no obvious ill effects.

II.Is Copyright Different?

              So far, we have argued that:

              1.  Copyright law restricts speech, but speech that copies another's expression is constitutionally unprotected.

              2.  In Bose, the Court mandated independent review of judgments that a particular instance of speech falls into a constitutionally unprotected category.

              3.  Therefore, independent review is required for judgments that a particular instance of speech copies another's expression.

              But might copyright law somehow differ from other speech restrictions (libel law, obscenity law, and the like), and thus deserve differing treatment?

A.The First Amendment Interest

              Perhaps copyright law is somehow specially immune from the normal concerns surrounding other speech restrictions, or perhaps copyright law vindicates especially important interests.  We have certainly heard this view among many, and especially among copyright lawyers.  Though copyright law is clearly a speech restriction, to many it lacks that speech restriction flavor.  It doesn't sound like censorship, only private people enforcing their lawful property rights.  Still, while many have this intuition, is there some specific reason underlying it, some reason that can justify setting aside the normal First Amendment procedural guarantees?

1.Property Rights

              The argument that copyright law should be immune from standard First Amendment procedural rules because it protects property rights strikes us as a non sequitur. 66  Free speech guarantees can't be avoided simply by characterizing a speech restriction as an "intellectual property law."  After all, one could plausibly view libel law as protecting a person's property interest in his reputation, or a company's property interest in its product's reputation. 67  One still has to ask -- as the Supreme Court has asked in its free speech/intellectual property cases 68 -- whether these intellectual property laws are unconstitutional speech restrictions, and, if they are substantively constitutional, what procedural protections the First Amendment nonetheless requires.

2.Private Enforcement

              Copyright law is largely enforced by private litigation, not government prosecution, so some argue that it's much less likely to turn into an engine of censorship.  But of course libel law is also largely enforced by private litigation.  Despite this, libel law is a government-imposed restriction, 69 even if the regime of private enforcement makes it a little harder for the government to use the restriction as part of a coherent censorship campaign. 70  Bose itself involved a private company suing over a statement that the government cared little about; nonetheless, the Court held that the First Amendment's procedural protections apply.  The same should go for copyright law.


              It's also incorrect to argue that intellectual property law is content-neutral and should therefore be subject to laxer rules. 71  To begin with, independent review is still required in cases involving content-neutral restrictions. 72  But beyond this, copyright liability turns on the content of what is published.  It's true that the law draws no ideological distinctions; just like libel laws, obscenity laws, and fighting words laws, it applies equally to speech advocating democracy, speech advocating Communism, and speech with no ideological message at all.  But while this might make the law viewpoint-neutral, it doesn't make it content-neutral, 73 and doesn't avoid the Bose rule.

4.Subject Matter of the Jeopardized Speech

              Some suggest that copyright cases pose less of a threat to free speech because they typically involve nonpolitical matters; after all, if a court erroneously concludes that Battlestar Galactica infringes the plot of Star Wars, 74 will the Republic really fall?  When the risk of error or chill falls only on such pedestrian material, there's no need for special procedural protections -- or so the argument would go.

              However, Bose itself was a trade libel case, involving nothing more important than a product review of a stereo speaker system.  The Court likewise applies independent review in obscenity and fighting words cases, 75 even though the risk of error or chill in such a case is typically borne by nonpolitical speech; 76 the same is true for commercial speech cases. 77  And of course quite a few copyright cases do involve political speech (consider Harper & Row itself), and quite a few libel cases involve material that seems to be as much "entertainment" as a typical movie or novel. 78

5.It's All Costs of Doing Business

              Many copyrights -- especially those copyrights that actually end up in litigation -- tend to be exploited by fairly large businesses.  One might argue that the chilling effect of legal uncertainty is less a problem when lawsuits are just a cost of doing business; Twentieth-Century Fox just isn't going to be very chilled by the risk of copyright litigation. 79

              But this in no way differentiates copyright from libel law.  Most libel defendants are also large businesses, whose job is reporting things, and whose business interests in many respects counteract any chilling effect the law could have.  And even wealthy entities that can afford a lawsuit might still be reluctant to face one; businesses try hard to minimize their costs of doing business.  That's why we worry that libel lawsuits might lead even the richest newspapers to soft-pedal certain issues.  The same goes for copyright cases:  A movie studio may be able to bear the cost of litigation over, say, a docudrama that's similar to another story based on the same set of facts, but this doesn't mean the studio will be willing to bear this cost.  It might decide to do another story instead -- perhaps no great loss to the world, but no less a loss than in the typical libel situation.

6.Copyright Law Furthers Free Speech Values

              Nor can copyright law be exempted from the general Bose rule on the grounds that "copyright itself [can] be the engine of free expression." 80  Copyright law's speech-enhancing effect, coupled with the specific constitutional authorization for copyright law, may justify holding copyright law to be a substantively valid speech restriction. 81  But the point of procedural rules -- such as the independent review principle -- is to make sure that even substantively valid speech restrictions don't end up restricting speech that should remain protected.

              Moreover, independent appellate review wouldn't in any event greatly diminish the incentive provided by copyright law, just as such review doesn't greatly diminish the force of libel law or obscenity law.  If anything, the premise that copyright law furthers free speech values makes it especially important that copyright cases be accurately decided, since error in either direction (too much protection or too little) implicates a First Amendment interest.  If Bose is correct -- if independent appellate review is important to correct unconstitutional results, and to refine the rules and thus reduce the risk of such errors in the future -- then independent review is doubly valuable here. 82

              We're also generally skeptical of distinguishing supposedly speech-furthering restrictions from other restrictions.  Many kinds of speech restrictions may be seen as furthering speech in some way.  Justice White argued this about libel law, claiming that "virtually unrestrained defamatory remarks about private citizens [may] discourage them from speaking out and concerning themselves with social problems." 83  Some have likewise argued that pornography tends to "silence" women, which might suggest that obscenity law may serve First Amendment values. 84  Justice Jackson argued that bans on strident public denunciations of a religion may serve First Amendment religious freedom values. 85

              Some contend that these arguments justify substantive speech restrictions and some contend the opposite.  But regardless of how one comes down on this, the arguments do not justify exemption from the normal procedural rules that make sure that the substantive rules, whatever they are, are accurately applied.  And in any event, copyright law's supposed speech-enhancing effects can't justify a special exemption for copyright law alone.

7.The Copyright Clause

              Unlike with libel and obscenity, the Constitution specifically refers to the government interest underlying copyright law.  Article I, section 8 of the Constitution authorizes Congress "[t]o promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries."  Harper & Row mentioned this as one reason why copyright law is a constitutionally permissible speech restriction. 86

              But the existence of the congressional power can't exempt copyright law from all First Amendment scrutiny.  The Copyright and Patent Clause grants power to Congress, but the point of the Bill of Rights is to restrain the federal government in the exercise of its enumerated powers.  In exercising the other enumerated powers, Congress is subject to First Amendment constraints; for instance, the government has the enumerated power to run the post office, but this doesn't mean it can refuse to carry Communist propaganda. 87  And in exercising its copyright power, Congress is bound by the Fourth, 88 Fifth, and Sixth Amendments.  Copyright law must likewise be bound by the First Amendment, too.

              We agree that it would be unsound to read the First Amendment as entirely eliminating the copyright power created by the Framers only two years earlier. 89  The Copyright and Patent Clause does represent the Framers' judgment that "copyright itself [can] be the engine of free expression," so courts ought not, in their zeal to protect speech, eviscerate the incentive that copyright law provides. 90  This was good reason for Harper & Row to conclude that copyright law is substantively constitutional.  But it hardly shows that copyright law ought to be free of the traditional procedural protections available in all other First Amendment cases. 91  Independent appellate review wouldn't eviscerate or even greatly diminish the incentive provided by copyright law, just as such it doesn't eviscerate libel law or obscenity law; it will merely require the law to be enforced in a slightly different way.

8.Importance of the Government Interest

              Finally, one might correctly point out that the interest upheld by copyright law -- the interest in providing an incentive for the dissemination of ideas -- is, even without regard to its constitutional status, quite important.  We do not doubt that this is so.

              But lots of speech restrictions are justified by important interests.  Preventing incitement to imminent violence is surely an important interest, too, and the Bose rule still applies to incitement cases.  Protecting individual reputation is an important interest. 92  So is combatting child pornography. 93  Independent review does not prevent these interests from being served; it just requires that they be served through a certain set of procedures.

B.Is Trying to Refine the Idea/Expression Dichotomy Pointless?

              The first foundation for Bose review -- the normative obligation to protect speech from erroneous restriction -- thus applies as well to copyright law as it does to libel law, obscenity law, and all the other areas where Bose review is required.  The heart of any assertion that Bose ought not apply to copyright cases must be a claim that the second justification -- that independent review will help the rule evolve through case-by-case adjudication -- is simply factually incorrect for copyright cases: that, as Learned Hand suggested, decisions about substantial similarity of expression "must . . . inevitably be ad hoc," and that the test must therefore be "of necessity vague." 94  Perhaps common law adjudication cannot give the line more meaning than the words themselves offer; perhaps ultimately each factfinder must draw the line anew, based on its own notions of what's "idea" and what's "expression."  Maybe the question must be determined in each case on its own facts, and new subrules and even new benchmark decisions are thus useless.

              We think there is a kernel of truth to this argument, but not enough to justify departing from Bose.  True, the idea/expression dichotomy is necessarily vague, and additional refinement by appellate decisions won't make it vastly clearer.  Still, the appellate decisionmaking that independent review makes possible will probably provide some extra clarity, some extra guidance for litigants and lower courts.  Coupled with the normative argument, this is enough to warrant following Bose even in copyright cases.

              We come to this conclusion based on four sources of evidence:

              1.  Copyright decisions do often compare and contrast the facts in a case with the facts in prior cases.  Decisions that determine whether two items are substantially similar in their expression -- rather than just deciding whether such a conclusion is clearly erroneous -- are thus of significant precedential value.

              2.  Lawyers likewise seem to rely in some measure on prior case law.

              3.  The Supreme Court has accepted that independent review is valuable even where, as here, the rule is fact-intensive and the precedent-building benefits of independent review are not likely to be tremendously large.  While the Court may have erred on this -- or perhaps even generally erred in Bose -- the Court's considered judgment in these cases deserves respect.

              4.  While there may be some costs to independent review, some circuits have adopted it (for reasons other than the ones that we suggest) with no obvious ill effects.  Likewise, all circuits generally independently review fair use findings, again with little trouble.

              Points 1 and 2 suggest that independent review may be practically beneficial.  Point 3 suggests that even a small benefit may be adequate to justify it.  Point 4 suggests that independent review is unlikely to be harmful.

              This evidence is tentative, and we're not sure how one could get more definitive evidence.  Still, given the Court's ruling in Bose, the burden of proof should be on those who want to carve out a copyright exception.  Absent such a showing, fidelity to the rules requires that Bose be followed; independent appellate review is therefore the sounder result.

1.Lower Courts' Use of Prior Cases as Benchmarks

              Despite the supposedly ad-hoc nature of each idea-expression decision, courts do look to benchmark cases in deciding whether two works are substantially similar in their expression.  Durham Industries, Inc. v. Tomy Corp., 95 a case dealing with alleged infringement of copyrights in games, dolls, and toys, is a good example.  The Durham court acknowledged that "it has been said that `[g]ood eyes and common sense may be as useful as deep study of . . . cases, which themselves are tied to highly particularized facts´"; 96 but, it went on to say, "we have nevertheless consulted the cases dealing with toys and dolls in order to check both our eyes and our sense." 97  The court then cited eleven cases on similar subjects, briefly comparing degrees of similarity; thus, for instance, the court acknowledged that plaintiff's and defendant's dolls were similar in that they walked or crawled, were small and made of plastic, and had "full faces, pert noses, bow lips, and large, widely spaced eyes," 98 but cited cases holding that these sorts of combinations of features are ideas, not expressions. 99

              Likewise, Conan Properties, Inc. v. Mattel, Inc. 100 relied on comparisons with earlier case law in holding that He-Man (a Masters of the Universe character) did not infringe the Conan the Barbarian character.  Saban Entertainment, Inc. v. 222 World Corp. 101 similarly quoted extensively from an earlier case in concluding that the Mega Rangers Power Bike set was substantially similar in its expression to the Mighty Morphin Power Rangers dolls.

              Moving from toys to compilations, in Key Publication, Inc. v. Chinatown Today Publishing Enterprises, Inc. 102 the Second Circuit reversed a lower court opinion, and held that while plaintiff's "Chinese Business Guide & Directory" was copyrightable, defendant's "Chinese American-Life Guide" did not copy the Business Guide's copyrightable selection and arrangement.  Judge Winters noted that the former selected 9000 businesses arranged in 260 categories, while the latter selected only 2000 businesses arranged in twenty-eight categories.  The selections and arrangements were not similar enough to constitute infringement.  The opinion drew on two earlier compilation cases, Eckes v. Card Prices Update 103 and Kregos v. Associated Press, 104 as well as the Supreme Court's decision in Feist Publications, Inc. v. Rural Telephone Service Co., Inc., 105 in deciding where to draw the line.

              Collectively, Eckes, Kregos, and Key Publications show the Second Circuit creating fairly detailed guideposts that help draw the idea/expression line in factual compilation cases.  Selecting 5000 out of 18,000 baseball cards as premium is copyrightable expression.  Choosing 9000 businesses for a business guide and arranging them into categories is expression, but creating another business guide with substantially different selections and arrangements takes only the idea, not the expression.  Putting together nine kinds of statistics on pitchers is just enough to qualify for protection, but the protection is limited to near exact copying.  Together these cases give some sense as to how much protection a compilation will likely receive.

              Even with literary works, courts use previous cases as guides and instruction (though this may at first glance seem harder, since literary works are much more complex than compilations or toys).  Consider Nash v. CBS, Inc. 106  Jay Robert Nash published many books claiming that John Dillinger wasn't actually killed in a shootout with FBI agents; rather, the theory went, a look-alike was shot in his place, and Dillinger laid low and survived for many decades.  An episode of the TV show "Simon and Simon" was based on this idea, and Nash sued for copyright infringement.  The District Court granted summary judgment for CBS, and the Seventh Circuit affirmed.

              In his opinion Judge Easterbrook points out that the existing tests for distinguishing idea from expression "do[] little to help resolve a given case":  "After 200 years of wrestling with copyright questions, it is unlikely that courts will come up with the answer any time soon, if indeed there is `an´ answer, which we doubt." 107  Easterbrook analyzes how providing either too little or too much protection may discourage speech, then says courts "must muddle through, using not a fixed rule but a sense of the consequences of moving dramatically in either direction." 108

              Despite this, the opinion does look to earlier cases as benchmarks:  It compares the fact pattern to that in Hoehling v. Universal City Studios, Inc., 109 which found no infringement because there was only substantial similarity of facts and not of expression, and Toksvig v. Bruce Publishing Co., 110 which found infringement because the expression was substantially similar.  The prior cases did create law that gave some guidance in a future case.  And the same happens in many other situations. 111

              Of course, these cases are hardly conclusive proof that independent appellate review is worth the candle.  Courts might just be citing the precedents as post hoc rationalizations of whatever view they held to begin with.  Or perhaps the precedents do provide useful benchmarks -- and are therefore useful predictors -- but only to a small degree; maybe each new precedent is helpful only in a few cases, and even there only to set a mood rather than to provide a definite answer.

              Nonetheless, the cases do undercut the strong claim that each copyright case is sui generis.  Judges at least assert that past decisions are helpful in deciding future ones.  Precedents do seem to "confine the perimeters" of the idea/expression dichotomy, to "mark[] out the limits of the standard through the process of case-by-case adjudication," to "give[] meaning" to the rule "through the evolutionary process of common law adjudication." 112  And the Court itself accepts the notion that litigation may help "the boundaries of copyright law [to] be demarcated as clearly as possible." 113

              The magnitude of this clarifying effect is unclear.  If Bose had never existed, and if the Court were deciding solely on pragmatic grounds whether to adopt independent appellate review in copyright cases, we might conclude that the clarifying effect was too speculative to justify independent review.

              But the Court has mandated independent review in Bose, both because of the supposed law-clarifying benefits of such review, and because of the perceived constitutional value of an extra look in cases involving free speech rights.  Given this, Bose must be followed unless there's strong reason to believe that its reasoning is mistaken in copyright cases.  The cases we cite above suggest that there is no such strong reason.

2.Lawyers' Use of Prior Cases as Benchmarks

              Another possible check on whether the Supreme Court's intuitions apply to copyright cases is the view of experienced lawyers.  Theory that sounds good in the ivory tower, or to the Justices, may end up failing on the streets.

              We are conscious of the great difficulty, perhaps even impossibility, of any scientific empirical work in this area.  Even if we could ask a large enough sample of lawyers their views on the question, we'd only get a sense of their beliefs (as filtered through the particular text of the questions we asked).  These beliefs might not quite match reality; they might not even reveal the lawyers' likely actions.  In fact, since many lawyers don't often consciously think about standards of review, they may not have any settled beliefs on the subject at all.

              Nonetheless, we talked to a handful of experienced lawyers to get a sense of their views -- a reality check against the judgment of several expert practitioners.  If, for instance, all of them had firmly told us that prior caselaw was useless in predicting future cases, this would have led us to think twice about our conclusions.

              We asked six copyright lawyers how they advise clients whether one work may have infringed another. 114  We started with general questions about how the lawyers predict what a court would do if presented with a specific fact pattern, and then focused on questions about appellate review. 115  Obviously, this is not a scientific survey of any sort -- the sample size is laughably small, and the questions are open-ended; but it may provide some however unreliable insight into how some lawyers react to cases.

              When we asked the lawyers generally how they would predict what a court would do -- without specifically raising the possibility of comparing this case against the precedents 116 -- we got a range of answers.  Some lawyers mentioned case law, among other things. 117  Others initially stressed that it is a matter of judgment. 118  Blaine Greenberg walks around the office with examples of the two items and asks non-lawyers what their reactions are. 119

              When asked explicitly whether they would look to the cases, 120 those who did not initially mention them split on their usefulness.  Bob Osterberg would turn to those cases he knows, and do research on cases outside his core field (music). 121  Herb Schwartz would look to cases, but "cases only take you so far in this area." 122  Greenberg looks to cases "only if we had a very specific area where I thought it had case law that goes to it," stressing the fact-specific nature of the question. 123

              The lawyers could think of few ways to make it easier to predict likely infringements. 124  Schwartz remarked that "it's like pornography, unfortunately." 125  They split over whether having more benchmark cases would help. 126  Tom Hemnes thinks that benchmark cases help in more established media, but that courts have trouble in more unusual areas. 127  Peter Nolan says "of course yes, if you have something very akin to your situation.  Unfortunately, courts sometimes differ in how they come out with these cases, even though they are applying the same rules." 128  Greenberg says "the more information that exists and the more published opinions there are, the better off you are.  But the more cases decided, the more likely it is that you can find a rationale for your argument because not all courts are going to agree." 129  Osterberg thinks more benchmarks might be useful in some areas, though he believes that in his area (literary property) there are plenty. 130  On the other hand, Schwartz doubts more benchmark cases would increase predictability. 131  David Nimmer thinks there are plenty of cases; indeed, "maybe it would make life easier if there were fewer benchmark cases." 132

              These responses reflect, sometimes simultaneously, two conflicting views of the cumulation of precedent.  One view holds that the buildup of precedents makes the law clearer, more detailed, and more predictable.  The other view sees case law becoming more chaotic with more precedents, as courts with different outlooks find ways to distinguish cases with close facts.

              Most of the lawyers do distinguish how they use district court and circuit court precedents. 133  This is partly because circuit cases are mandatory precedent, but there are other reasons as well.  Greenberg feels that district courts are less predictable than circuit courts.  "Trial judges are such strong personalities that they don't care what their brethren will do.  It wouldn't surprise me if some cases would go ten different ways depending on who you draw. . . .  But if the case can go to the Ninth Circuit, and if the Circuit is consistent, I may anticipate winning at the appellate level." 134  As for opinions from other circuits, the lawyers differ.  Osterberg says that even cases from other circuit courts "get more attention and respect" than do district court cases. 135  Nolan had a similar opinion, noting that, among other things, circuit cases involve three judges.  Nimmer and Schwartz were less inclined to pay attention to cases from other circuits.

              When asked whether they treat circuit court cases reviewing summary judgments differently from those reviewing the results of trials differently, 136 the lawyers gave varying answers.  Nimmer treats them differently "only to predict if this is a case that's right for summary judgment." 137  Greenberg agrees, and says that if a case survives summary judgment and gets to trial, "then I would advise the client that I think the equitable appeal of the case will be a stronger determinant than the case law." 138  Schwartz takes a view closer to ours, saying "denials of summary judgment don't tell you a lot.  Granting summary judgment is a pretty extreme case.  So trials tell you more." 139

              On the points most closely tied to this article, most of the lawyers do not pay attention to whether circuit court cases used de novo review or clear error review. 140  However, Hemnes says "if the court holds a decision is not clearly erroneous but has dicta saying it is questionable and it might hold differently under a different posture, then I treat clear error cases differently.  Usually the court will make that clear." 141  And Osterberg does say "I would tend to give greater weight to decisions using de novo review." 142  When asked whether they were aware of differences among circuits in the standard of review, 143 some had some sense of a difference, and some did not, but none had any detailed response as to the differences.  When asked about general differences among circuits, 144 Hemnes said "the circuits that address this comparatively infrequently are harder to predict than those which have a long track record.  At the most extreme, the Second Circuit is pretty predictable. . .  On the other hand, the Ninth Circuit is a puzzle to me.  Youžd think they would have a lot of experience, but they are less predictable to me.  It may be that they are struggling with less traditional media." 145

              Some of the lawyers did think that the standard of review the court uses might affect their decision whether or not to appeal a trial court's holding. 146  Greenberg says "if there is de novo review, you've got a much better chance of winning.  If there is de novo review, we'd be more likely to appeal, and more likely to be interested in what the court of appeals had done on these cases." 147  Nimmer says he "should be" influenced by the standard of review:  "If the standard is de novo, and I think it's a good case for my client, I should appeal.  If it is clear error and the judge preserved the record, then I'm sunk." 148

              Not all, however, thought the standard of review would affect their decision whether to appeal.  Schwartz denied it would have an effect.  Nolan said "if it were de novo I'm more likely to go up.  But unless a pile of money is involved we would not appeal, we would settle.  It's too costly to appeal."  He uses appeals mainly to establish legal principles. 149

              Thus, our informal inquiry suggests that these practicing lawyers have not particularly considered the question of the effect of appellate review standards on which we focus in this paper.  Some of them share the view that in copyright law the cases are quite fact-specific, and good judgment rather than specific case law is a lawyer's most important guide.  But at least some of them think that more benchmark cases, particularly on close fact patterns, would make outcomes more predictable, and they generally agree that circuit court opinions are more useful than trial court opinions.  The judgments of these expert practitioners thus reinforce our view that devices which force circuit courts to produce more detailed, useful opinions could be of some help.  De novo appellate review is one such device.

3.The Supreme Court's Views in Other Areas

              Case law and the interviews with lawyers thus suggest that extra precedents may help, to a limited extent.  Is this limited benefit sufficient to justify independent review?

              The Court has recently confronted this question in other constitutional contexts, and has generally come down in favor of independent review even when its law-clarifying benefits are modest.  The most recent case, Ornelas v. United States, 150 involved review of reasonable suspicion and probable cause conclusions in Fourth Amendment cases.  These are fact-intensive, totality-of-the-circumstances inquiries, and the Court began by stressing that

              Articulating precisely what "reasonable suspicion" and "probable cause" mean is not possible.  They are commonsense, non-technical conceptions that deal with "`the factual and practical considerations of everyday life on which reasonable and prudent men, not legal technicians, act.´" As such, the standards are "not readily, or even usefully, reduced to a neat set of legal rules." . . .  They are . . . fluid concepts that take their substantive content from the particular contexts in which the standards are being assessed. 151 This strongly resembles what some say about idea/expression judgments.

              Despite this, the Court held that independent appellate review was required: [T]he legal rules for probable cause and reasonable suspicion acquire content only through application.  Independent review is therefore necessary if appellate courts are to maintain control of, and to clarify[,] the legal principles. . . .  [W]here the "relevant legal principle can be given meaning only through its application to the particular circumstances of a case, the Court has been reluctant to give the trier of fact's conclusions presumptive force and, in so doing, strip a federal appellate court of its primary function as an expositor of law[.]"

              Finally, de novo review tends to unify precedent and will come closer to providing law enforcement officers with a defined "`set of rules which, in most instances, makes it possible to reach a correct determination beforehand as to whether an invasion of privacy is justified in the interest of law enforcement.´" . . .  "[T]he law declaration aspect of independent review potentially may guide police, unify precedent, and stabilize the law" . . . .

              It is true that because the mosaic which is analyzed for a reasonable-suspicion or probable-cause inquiry is multi-faceted, "one determination will seldom be a useful `precedent´ for another."  But there are exceptions [giving examples].  And even where one case may not squarely control another one, the two decisions when viewed together may usefully add to the body of law on the subject. 152

              The Court took a similar approach in Thompson v. Keohane, 153 where it confronted the proper standard of review on habeas corpus of state court determinations whether a person was "in custody" for Miranda v. Arizona purposes.  It acknowledged that the inquiry had a "`totality of the circumstances' cast," 154 and that appellate decisions might not be able to "supply `a definite rule.´" 155  Nonetheless, it held that "`in custody´ determinations do guide future decisions"; that though they can't create definite rules, "they nonetheless can reduce the area of uncertainty"; 156 and that they may "unify precedent[] and stabilize the law." 157

              When no constitutional values are at stake, 158 or when the application of a test is generally interwoven with witness credibility determinations 159 -- which is generally not the case in idea/expression judgments, which often require only the comparison of physical items 160 -- deferential review may be proper.  But as Ornelas and Thompson suggest, where constitutional principles presumptively require independent judgment review, courts shouldn't be stymied by the fact that the test "[can]not readily, or even usefully, [be] reduced to a neat set of legal rules." 161

4.Independent Appellate Review in Those Circuits That Already Do It

              The Second Circuit already independently reviews findings of substantial similarity of expression.  There is little sign of this causing any great problems; we know of no commentaries or cases pointing to specific bad results of such review, or even generally bemoaning its presence.

              Our tentative impression, in fact, is that the Second Circuit has on the whole produced a more helpful body of caselaw than the other big copyright circuit, the Ninth. 162  This is only a tentative impression, and even if it's correct, there may be other explanations:  The Ninth Circuit has more judges and may thus produce less consistent caselaw; the West Coast also seems to produce more cutting-edge new media cases, where it might be harder to create clear, settled precedent.  Nonetheless, so long as the Second Circuit does no worse than the Ninth, the main argument against following Bose -- that the practical costs are so great that we should carve out an exception to the precedent -- falls away.

              Moreover, all circuits are already familiar with independent appellate review in at least one area of copyright law: fair use. 163  Fair use, like substantial similarity of expression, has generally been seen as extremely fact-intensive, "requir[ing] a difficult case-by-case balancing of complex factors." 164  Despite this, though, appellate decisions have created some greater predictability.  In the words of Justice Kennedy, "[t]he common-law method instated by the fair use provision of the copyright statute . . . presumes that rules will emerge from the course of decisions," 165 and it seems that they have.  Following Sony Corp. v. Universal City Studios, 166 we know that certain sorts of private noncommercial copying of entire works are generally permissible.  Following Campbell v. Acuff-Rose Music, 167 we know that parodies commenting on the work from which they borrow are favored uses akin to commentary or news reporting.  "[T]he evolutionary process of common law adjudication" 168 seems to work for fair use, which leads us to suspect that it might also work (or at least not cause enough harm to justify disregarding Bose) for substantial similarity of expression.

C.Against Special Pleading for Copyright

              We thus see no compelling normative reason to treat copyright differently from other speech restrictions, restrictions that are likewise substantively valid but nonetheless require certain procedural safeguards.  There's some practical reason to be skeptical of Bose's utility, but the arguments against applying Bose aren't enough to avoid its precedential force.

              And we see a good reason against treating copyright better than other speech restrictions.  The First Amendment demands sacrifices from many who earnestly believe in the legitimacy of their favorite speech restrictions.  This burden is always heavy, but it becomes heavier still when others' speech restrictions -- especially restrictions that, like copyright law, are identified with the relatively rich and powerful -- are being given a free ride. 169  Partisans of hate speech restrictions have already begun to point this out; while we disagree with their substantive proposals, they are right to demand that people who oppose proposed new exceptions to First Amendment defend their support for the old exceptions.

              Any special preference for copyright law must thus be justified by some pretty substantial difference between copyright and other speech restrictions.  Where, as here, no such difference exists, favoritism for a particular kind of speech restriction risks corroding public respect for First Amendment law more generally.

III.Reconsidering Bose?

              Thus, Bose constitutionally requires independent judgment review of idea/expression conclusions; there's no good justification for carving out a special exception for copyright cases.  And yet we suspect that quite a few readers, especially copyright lawyers, may still conclude that the majority of circuits is right, and that independent review is wrong.  Independent review does impose a burden, both on litigants and the appellate system:  It will require appellate courts to spend more time on each appeal, 170 and it may encourage litigants to appeal, 171 thus delaying the final decision and consuming more court time and litigant money.  And the benefits that it provides seem somewhat speculative.

              This, it seems to us, provides a good opportunity to ask whether Bose itself might be mistaken.  We won't engage in a complete analysis here, but we'll briefly sketch how such an argument might go, and what it would mean for First Amendment law more broadly.

              Any system of justice administered by humans has necessary imperfections.  Factfinders might be biased or simply wrong.  Rules will not be perfectly well-defined, which increases the risk of factfinder bias in applying the rules, and leads cautious people to "steer far wider of the unlawful zone." 172

              Normal law reacts with some equanimity to these imperfections.  Sure, judges and juries will sometimes get matters wrong, but that's life.  Sure, the law is vague, but we cannot demand perfect clarity in our complex world.  Perhaps the vagueness creates something of a chilling effect, but the chilling effect is unavoidable, tolerable, and not really that great; people are, after all, used to uncertainty and risk, and are unlikely to overreact to it.  Life is uncertain, and the law, to use a phrase one of us has often heard from his father, "is as vague as life itself." 173  Might as well deal with it.

              One could easily imagine a constitutional jurisprudence that imposes some substantive restraints -- for instance, a requirement that libel law allow truth as a defense, 174 or even a requirement that public figures suing over statements of public concern prove "actual malice" 175 -- but then leaves the law largely alone.  Such a system wouldn't trouble itself over independent review; mistakes happen, rules are vague, but that's too bad.  Nor would such a system put restrictions on preliminary injunctions, on quanta of proof, on burdens of proof, or what have you.

              This approach appears in some of Justice White's libel opinions.  White consistently endorsed New York Times v. Sullivan, 176 but was skeptical of the further constitutionalization of libel law.  He was, for instance, unpersuaded that strict liability for falsehoods about private citizens was a "realistic threat to the press and its service to the public"; 177 this risk was to him subordinate to the normal tort law principle of compensating innocent victims for the deliberate, even if honestly mistaken, actions of a publisher. 178  He did not "fear[] uncontrolled awards of damages by juries," because of his confidence in "the good sense of most jurors" and in the (quite limited) normal-law appellate review mechanisms that would set aside the obviously unreasonable verdicts. 179  As to independent review, Justice White largely joined Justice Rehnquist's dissent in Bose that suggested -- among other things -- that "the need to shield protected speech from the risk of erroneous fact-finding" 180 doesn't require independent review of certain issues.

              The alternative to this view is the "pervasive constitutional concern" model: a mistrust of the normal law system, and a concern that its imperfections are intolerable in the First Amendment context.  Under this view, the risk of factfinder errors, as errors of constitutional dimension, necessitates extra review procedures.  The risk of constitutional violations must be minimized, even at the expense of increasing the risk of uncompensated harm.  The chilling effects of substantial liability and of vague laws are presumed, even without empirical evidence of their magnitude.  Constitutional freedoms are seen as "fragile" 181 and in need of special procedural protection; "rigorous procedural safeguards" are needed because "the freedoms of expression must be ringed about with adequate bulwarks." 182

              The tension between these views is clearly visible in First Amendment jurisprudence, especially libel jurisprudence.  It is likewise visible in other areas, such as death penalty cases, where the Court, not content with just deciding whether the death penalty is substantively constitutional, has concluded that the danger of normal law's imperfections is so great as to justify additional procedural requirements.

              The Bose rule rejects the normal law model.  Normally, Bose acknowledges, Rule 52(a) requires deference to lower courts (at least as to pure questions of fact) and the Seventh Amendment requires similar deference to juries; but where constitutional liberties are involved, the rule must be different.  The judgment behind this is partly empirical -- courts of appeals will correct factfinder mistakes and refine vague rules through case-by-case adjudication -- but primarily normative:  Courts of appeals have a constitutional duty to try to correct factfinder mistakes and to try to refine vague rules.  The normative value of such attempts at correction and refinement, even if they aren't always practically very helpful, justifies the tangible costs to the judicial system and the litigants.

              The Court's arguments for adopting the pervasive constitutional concern model strike us as persuasive.  Writers who ask "What may I lawfully write?" are entitled to a clearer and better policed standard than manufacturers who ask "What safety devices does the duty of `reasonableness' require me to implement?" 183  Free speech, as a constitutional right, deserves special protection.

              But we confess that the experience of copyright law tests those arguments.  Copyright law has led a largely normal-law existence, unencumbered with the various procedures the Court has demanded in libel cases, but the skies haven't fallen.  There might be a chilling effect here and there, some people steering "far wider of the unlawful zone," 184 and the occasional (or even frequent) unjust verdict, but writers keep writing and the marketplace of ideas doesn't seem strikingly impoverished.

              As Part II explains, we believe that none of this justifies treating copyright law differently from other restrictions.  Nonetheless, it does suggest some skepticism about the Court's insistence on the pervasive constitutional concern model in free speech cases generally, and might perhaps even counsel some rethinking of the model.

              One purpose of our article has been to show that, if one believes that independent review ought not be required in copyright cases, the only principled way to avoid this requirement is to reconsider it on a general basis.  If the Court has indeed gone too far in its distrust of the normal law adjudication process, all those harmed by speech -- whether speech that infringes on a copyright, speech that unjustifiably defames, or speech that counsels violence -- should be entitled to the benefits of any retrenchment that might be proper.

IV.  First Amendment Due Prcoess and Copyright Law

              Finally, we also hope to lead people to think more generally about "First Amendment Due Process" in copyright cases.  Harper & Row upheld copyright law substantively, but that itself tells us little:  Libel law, for instance, is also substantively constitutional, but the Court has imposed many procedural requirements for libel cases.  While these need not necessarily apply equally to all other speech restrictions, one can credibly argue that they should be so applied.

              Thus, we suggest several areas for future research -- and litigation:

              1.  Prior Restraint:  Temporary restrictions on speech justified merely by the possibility that the speech might be unprotected -- as opposed to a judicial finding that the speech is unprotected -- are generally considered unconstitutional prior restraints. 185  Are preliminary injunctions in copyright cases, which can be based merely on a finding of likelihood of success on the merits, similarly unconstitutional?  If not, what does this tell us about the propriety of similar preliminary injunctions in libel cases and other cases where important private rights may be jeopardized by continued misbehavior pending final judgment? 186

              2.  Strict Liability:  Laws that impose strict liability for certain kinds of speech have generally been struck down on the grounds that they would lead to too much self-censorship on the part of cautious speakers and distributors.  Thus, reasonable mistake of fact is a defense in public concern libel cases, even when the plaintiff is a private figure. 187  Reasonable ignorance of the content of the material that one is distributing is a defense in obscenity cases; 188 reasonable mistake as to the age of a performer is a defense in child pornography cases. 189  Is copyright law's imposition of strict liability for infringement thus wholly or partly unconstitutional? 190

              3.  Punitive and Presumed Damages:  In public concern libel cases, punitive damages and presumed damages may only be awarded against defendants who told falsehoods knowingly or recklessly. 191  In copyright cases, though, statutory damages -- which are the rough equivalent of libel's presumed (and perhaps even punitive) damages -- may be awarded even against negligent or innocent defendants. 192  Is this unconstitutional?  Does the cap of $20,000 per work infringed save the law from unconstitutionality, and if it does, does this suggest that similarly modest presumed damages in libel cases should be allowed even without a finding of "actual malice"?

              4.  Burden of Proof:  In public concern libel cases, the plaintiff bears the burden of proving falsehood, though under the common law the defendant had the burden of proving truth. 193  Harper & Row strongly suggested that speech that constitutes fair use, like speech that is defamatory but true, is constitutionally protected. 194  Does it follow that the burden of proving unfair use must likewise be put on the plaintiff? 195

              5.  Quantum of Proof:  In public figure libel cases, the plaintiff must prove actual malice by clear and convincing evidence, not just by a preponderance of the evidence. 196  Some lower courts impose a similar standard on proof of obscenity in civil obscenity injunction cases. 197  Should a similar standard be required for proof of substantial similarity of expression, or of unfair use? 198

              The answer to all these questions may be that copyright is different, and that all First Amendment bets are therefore off.  But we doubt that this is so, and we believe that the issues are certainly contestable enough to merit further investigation, and assertion by lawyers.


              If the Bose reasoning is right -- and we think it probably is -- then independent appellate review should help make copyright law somewhat clearer and more predictable.  The "substantial similarity" test cries out for the elaboration that independent appellate review is generally thought to provide.  We can't promise dramatic improvements, but we think that, where speech restrictions are concerned, even modest clarification will help.

              But in any event, whether or not the Bose reasoning is right (and we agree that the matter is not free from doubt), Bose is the law, and we see no good justification for courts to ignore Bose's command here. 199  Copyright law may be a permissible speech restriction, but it can't claim a categorical exemption from all free speech concerns.  The same First Amendment Due Process rules that apply to other substantively valid restrictions must, absent a very good reason, apply equally to copyright.  



              Following are the questions asked of practitioners, the answers to which are reported in Part II.B.2.

              Say a client comes to you and says he wants to publish something, but he's afraid of a copyright lawsuit.  It might be a book or a song or a movie or a factual compilation.  He knows he's been influenced by a preexisting work; he thinks he's only taken the idea; but he doesn't know how a court would see things.

              1.  How do you decide whether he's borrowed only the idea or the expression?  How do you predict what a court would do if it came to court?

              2.  (If cases not mentioned:)  Do you also look at the cases, and compare your situation against cases that have found substantial similarity and cases that haven't found it?

              3.  (If the answer to this is no:)  How do you give a relatively precise answer?

              4.  What, if anything, do you think might be done to make it easier for you to decide what's likely an infringement and what isn't?

              5.  Do you think it would be easier for you to predict how courts will behave if there were more benchmark cases against which to compare your situation?

              6.  In using past cases to predict results, do you distinguish between how you use Circuit and District Court cases?  If so, how?

              7.  In using past cases to predict results, do you distinguish Circuit Court cases reviewing summary judgments from Circuit Court cases reviewing the results of trials?

              8.  In using past cases to predict results, do you distinguish Circuit Court cases that have reached a decision using de novo review from cases that have reached a decision using clear error review?

              9.  Say that, in a bench trial, the judge concludes that your client's work was substantially similar to the plaintiff's.  Would your decision about whether to appeal be influenced by the standard of review on appeal (de novo as opposed to clear error)?

              10.  Have you noticed any differences between Circuits in the way they deal with cases?  (If yes:)  What differences?

              11.  Are you aware of different standards of review -- reviewing findings of substantial similarity de novo as opposed to for clear error -- between Circuits?  (If yes:)  Do the different standards of review affect the overall case law?  


    *    Acting Professor, UCLA Law School (  Many thanks to Zoe Hilden, Lionel Sobel, and John Wiley for their help, and to the John M. Olin Foundation for their extremely generous research assistance.

    **    Law Clerk to Judge Alex Kozinski, U.S. Court of Appeals for the Ninth Circuit.

    1.    Note that we speak here of copyright claims brought based on material that's "expression" for free speech purposes -- books, movies, songs, paintings, and so on.  Our argument doesn't cover copyrighted software, which (at least in object code) generally doesn't qualify as speech for First Amendment purposes.

    2.    Harper & Row Publishers, Inc. v. Nation Enterprises, 471 U.S. 539 (1985).

    3.    466 U.S. 485 (1984).

    4.    Id. at 505.

    5.    See infra Part I.E.

    6.    Id. at 502.

    7.    We discuss here substantial similarity, rather than probative similarity.  Copyright law prohibits (1) copying of (2) another's expression.  Courts consider whether the plaintiff's and defendant's works are similar for both prongs of this inquiry.  For the first prong, they ask whether the similarity is probative of the fact of copying; for the second, they ask whether the similarity is substantial enough to make the defendant's action into copying of the expression and not just of the idea.  See Paul Goldstein, 2 Copyright § 8.0; Melville B. Nimmer & David Nimmer, Nimmer on Copyright, § 13-28 n.3.2; Latman, "Probative Similarity" As Proof of Copying:  Toward Dispelling Some Myths in Copyright Infringement, 90 Colum. L. Rev. 1187, 1190 (1990).  Probative similarity is relevant to a purely factual question: did copying take place?  Substantial similarity is a question of degree, and of application of law to fact.

              We also do not discuss any glosses on substantial similarity that have been implemented by circuit courts, such as the Ninth Circuit's bifurcated intrinsic/extrinsic test.  See Shaw v. Lindheim, 919 F.2d 1353 (9th Cir. 1990).  Our points apply to substantial similarity of expression in all circuits.

              Finally, we don't generally discuss independent review of fair use questions, because courts already conduct such review, at least when the factfinder below was a trial judge.  See infra Part II.B.4; Harper & Row Publishers, Inc. v. Nation Enterprises, 471 U.S. 539, 560 (1985); Maxtone-Graham v. Burtchaell, 803 F.2d 1253, 1258-59 (2nd Cir. 1986) (holding that fair use is for the trial judge to determine at summary judgment, so long as underlying historical facts are not in dispute).  The points we make in this article might, however, be relevant to cases involving review of jury verdicts.

    8.    See Henry P. Monaghan, First Amendment "Due Process," 83 Harv. L. Rev. 518 (1970).

    9.    See infra notes 170 and 171 and accompanying text.

    10.    Cf. Spence v. Washington, 418 U.S. 405, 417 (1974) (Rehnquist, J., dissenting) (giving copyright law as an example of a speech restriction).

    11.    17 U.S.C. § 506 (criminalizing certain kinds of infringement).

    12.    E.g., Tin Pan Apple Inc. v. Miller Brewing Co., 30 U.S.P.Q.2d (BNA) 1791 (S.D.N.Y. 1994).

    13.    E.g., Steinberg v. Columbia Pictures Industries, 663 F. Supp. 706 (S.D.N.Y. 1987); Kisch v. Ammirati & Puris, Inc., 657 F. Supp. 380 (S.D.N.Y. 1987); cf. Woods v. Universal City Studios, Inc., 920 F. Supp. 62 (S.D.N.Y. 1996) (enjoining distribution of the movie 12 Monkeys because one scene infringed on a copyrighted drawing).

    14.    E.g., Toksvig v. Bruce Publishing Co., 181 F.2d 664 (7th Cir. 1950); Sheldon v. Metro-Goldwyn Pictures Corp., 81 F.2d 49 (2d Cir. 1936); cf. Judge Refuses to Halt Release of `Amistad,´ Rejects Writer's Claim, Wall. St. J., Dec. 9, 1997, at B16 (describing lawsuit by writer who claimed that Steven Spielberg's movie Amistad was based on her novel about the same historical incident; the court concluded that there was likely no infringement in this particular case).

    15.    Harper & Row Publishers, Inc. v. Nation Enterprises, 471 U.S. 539 (1985).

    16.    403 U.S. 713, 726 n.* (1971) (Brennan, J., concurring).

    17.    Accord Zacchini v. Scripps-Howard Broad. Co., 433 U.S. 562, 577 n.13 (1977) (some citations omitted): We note that Federal District Courts have rejected First Amendment challenges to the federal copyright law on the ground that "no restraint (has been) placed on the use of an idea or concept."  See also Walt Disney Productions v. Air Pirates, 345 F. Supp. 108, 115-116 (N.D. Cal. 1972) (citing Nimmer, Does Copyright Abridge The First Amendment Guarantees of Free Speech and Press?, 17 UCLA Rev. 1180 (1970), who argues that copyright law does not abridge the First Amendment because it does not restrain the communication of ideas or concepts). See also Lee v. Runge, 404 U.S. 887 (Douglas, J., dissenting from denial of certiorari) ("Serious First Amendment questions would be raised if Congress' power over copyrights were construed to include the power to grant monopolies over certain ideas."); Toro Co. v. R & R Products Co., 787 F.2d 1708 (8th Cir. 1986) (the idea-expression dichotomy is partly grounded "in the First Amendment interest in the free exchange of ideas," citing Harper & Row); Walt Disney Productions v. Air Pirates, 581 F.2d 751 (9th Cir. 1978) ("the idea-expression line represents an acceptable definitional balance as between copyright and free speech interests"); Sid & Marty Krofft Television Productions, Inc. v. McDonald's Corp., 562 F.2d 1157 (9th Cir. 1977) ("idea-expression dichotomy already serves to accommodate the competing interests of copyright and the first amendment").

    18.    See, e.g., Melvile Nimmer, Does Copyright Abridge the First Amendment Guarantees of Free Speech and Press?, 17 UCLA L. Rev. 1180 (1970).

    19.    Curtis Butts Publishing Co. v. Associated Press, 388 U.S. 130, 155 (1967); New York Times Co. v. Sullivan, 367 U.S. 254, 279-80 (1964).

              Portions of the analysis in this section are borrowed from Eugene Volokh, Freedom of Speech and Appellate Review in Workplace Harassment Cases, 90 Nw. U. L. Rev. 1009 (1996).

    20.    Bose Corp. v. Consumers Union, 466 U.S. 485, 505 (1984).  Professor Henry Monaghan describes this quote and the text quoted infra at note as the "core of the [Bose] opinion."  Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L. Rev. 229, 243 (1985).

    21.    466 U.S. at 511.  "The principle of viewpoint neutrality that underlies the First Amendment itself also imposes a special responsibility on judges whenever it is claimed that a particular communication is unprotected."  Id. at 504 (citation omitted).

    22.    Id. at 504.

    23.    Id. at 502.

    24.    Id.

    25.    Id. at 511.

    26.    Cf., e.g., MCA, Inc. v. Wilson, 677 F.2d 180 (2nd Cir. 1981) (concluding that the copyright infringement holding below wasn't clearly erroneous, and that neither would have been a contrary holding).

    27.    Baggett v. Bullitt, 377 U.S. 360, 372 (1964) (citation omitted).  Cf. Neil W. Netanel, 106 Yale L.J. 283, 303-04 (1996) ("the copyright law safeguards that have made First Amendment defenses seem overly intrusive and unnecessary have in fact been only sporadically effective in protecting First Amendment values," in part because "while the idea/expression dichotomy makes sense in principle, it is notoriously malleable and indeterminate"); id. at 381 (describing how "prevailing uncertainties" in copyright law interfere with the creation of certain kinds of new works); Diane Leenheer Zimmerman, Information as Speech, Information as Goods, 33 Wm. & Mary L. Rev. 665, 709 (1992) ("the fuzzed line between idea and expression" creates "uncertainty [that] can cast a serious chill on communicative activities"); Alfred C. Yen, A First Amendment Perspective on the Idea/Expression Dichotomy and Copyright in a Work's `Total Concept and Feel', 38 Emory L.J. 393 (1989) (generally stressing how the vagueness of the idea/expression dichotomy can deter constitutionally protected speech).  See generally Jessica Litman, Reforming Information Law in Copyright's Image, 22 U. Dayton L. Rev. 587, 600-02 (1997).

    28.    Bose, 466 U.S. at 503.

    29.    Id. at 504-08 (citing Supreme Court cases that apply this rule to fighting words, incitement, obscenity, and child pornography); see also Hurley v. Irish-American Gay, Lesbian & Bisexual Group, 115 S. Ct. 2338, 2344 (1995) (applying Bose to the question whether conduct was expressive); Miller v. Fenton, 474 U.S. 104, 114 (1985) (applying Bose by analogy in the Due Process Clause context to the question whether confession was voluntary); Murphy v. I.S.K. Con., 571 N.E.2d 340, 345 (Mass. 1991) (reading Bose as applicable to Free Exercise Clause issues); New Life Baptist Church Academy v. Town of East Longmeadow, 885 F.2d 940, 941-42 (1st Cir. 1989) (same).  See also Thompson v. Keohane, 116 S. Ct. 457 (1995) (applying similar analysis to in-custody determinations for Miranda purposes), and Ornelas v. United States, 116 S. Ct. 1657 (1996) (applying similar analysis to probable cause decisions in Fourth Amendment cases), both discussed infra in Part II.B.3.

    30.    E.g., Luke Records, Inc. v. Navarro, 960 F.2d 134, 138 (11th Cir. 1992).

    31.    E.g., Herceg v. Hustler Magazine, Inc., 814 F.2d 1017, 1021 (5th Cir. 1987); Yakubowicz v. Paramount Pictures Corp., 536 N.E.2d 1067, 1071 (Mass. 1989).

    32.    E.g., Braun v. Soldier of Fortune Magazine, Inc., 968 F.2d 1110, 1120-21 (11th Cir. 1992).

    33.    E.g., Standing Comm. on Discipline v. Yagman, 55 F.3d 1430, 1443 (9th Cir. 1995).

    34.    E.g., Swineford v. Snyder County, 15 F.3d 1258, 1265 (3d Cir. 1994); Mekss v. Wyoming Girls' Sch., 813 P.2d 185, 194 (Wyo. 1991); see also Rankin v. McPherson, 483 U.S. 378, 385-87 nn.8-9 (1987) (arguably dictum).

    35.    E.g., AIDS Action Comm. v. Massachusetts Bay Transp. Auth., 42 F.3d 1, 6-7 (1st Cir. 1994); see also Brown v. Palmer, 915 F.2d 1435, 1441 (10th Cir. 1990) (independently reviewing factual findings underlying the determination of whether a forum is public), aff'd, 944 F.2d 732 (10th Cir. 1991) (en banc).

    36.    E.g., Peel v. Attorney Registration and Disciplinary Comm'n, 496 U.S. 91, 108 (1990) (plurality); id. at 111-17 (Marshall, J., concurring) (engaging in independent review, but not citing Bose directly); Joe Conte Toyota, Inc. v. Louisiana Motor Vehicle Comm'n, 24 F.3d 754, 755-56 (5th Cir. 1994); Don's Porta Signs, Inc. v. City of Clearwater, 829 F.2d 1051, 1053-54 & n.9 (11th Cir. 1987).

              Some cases reviewing federal administrative agency findings don't appear to follow Bose.  Two such cases involve review of FTC findings that ads are false or misleading.  Another line of cases (including a Supreme Court case) involves review of NLRB findings that unionization-related speech by an employer or a union was impermissibly coercive.  Both lines are grounded on a deference-to-expert-agencies rationale.

              The FTC cases are Kraft, Inc. v. FTC, 970 F.2d 311, 316-17 (7th Cir. 1992), and FTC v. Brown & Williamson Tobacco Corp., 778 F.2d 35, 41 n.3 (D.C. Cir. 1985).  (Both cases also argued that Bose was inapplicable to commercial speech, but that seems to be in considerable tension with the Supreme Court's position in Peel, as well as the circuit decisions in Joe Conte Toyota and Don's Porta Signs.)  Cf. Martin H. Redish, Product Health Claims and the First Amendment, 43 Vand. L. Rev. 1433, 1459-60 & n.144 (1990) (criticizing Brown & Williamson on Bose grounds).  The NLRB cases follow NLRB v. Gissel Packing Co., 395 U.S. 575, 620 (1969), which held that "a reviewing court must recognize the Board's competence in the first instance to judge the impact of utterances made in the context of the employer-employee relationship."  Since Gissel, lower courts have not applied independent judgment in this area but have instead reviewed NLRB findings for "substantial evidence."  See, e.g., DTR Indus., Inc. v. NLRB, 39 F.3d 106, 114 (6th Cir. 1994).

              Gissel came long before Bose, and no court has confronted the tension between them, though distinguished commentators have pointed to the discrepancy.  Richard H. Fallon, Jr., Of Legislative Courts, Administrative Agencies, and Article III, 101 Harv. L. Rev. 915, 976, 990 (1988); Monaghan, supra note 20, at 244 & n.84, 258.  In any event, to our knowledge no court or commentator has suggested that substantial evidence review be transplanted from the expert agency setting to the review of findings made by judges and juries, where Bose is firmly entrenched.

              The only non-agency case we could find that declined to follow Bose in determining whether speech is unprotected, Levine v. CMP Publications, Inc., 738 F.2d 660 (5th Cir. 1984), seems to no longer be good law.  Levine involved the finding that defamatory statements about private figures were made negligently.  The court reasoned that Gertz v. Robert Welch, Inc., 418 U.S. 323 (1974), "allow[ed] the states to regulate [private figure defamation] within much less restrictive bounds than those imposed [on public figure defamation]," and that therefore Bose was inapplicable.  Levine, 738 F.2d at 672 n.19.  But after Levine was decided, the Court made clear that Bose does indeed apply to negligence findings in private figure cases.  See Milkovich v. Lorain Journal Co., 497 U.S. 1, 20 (1990); see also LeDoux v. Northwest Publishing, Inc., 521 N.W.2d 59, 69 (Minn. Ct. App. 1994) (applying Bose in such a situation); Turf Lawnmower Repair, Inc. v. Bergen Record Corp., 655 A.2d 417, 423 (N.J. 1994) (same).

    37.    Turner Broadcasting System, Inc. v. FCC, 114 S. Ct. 2445, 2471 (1994); Association of Community Org. for Reform Now v. St. Louis County, 930 F.2d 591, 595-96 (8th Cir. 1991).

    38.    Bose Corp. v. Consumers Union, 466 U.S. 485, 503-06 (1984).  And refining the definition of "substantial similarity of expression" serves the goals of copyright law as well as of the First Amendment:  "Because copyright law ultimately serves the purpose of enriching the general public through access to creative works, it is peculiarly important that the boundaries of copyright law be demarcated as clearly as possible."  Fogerty v. Fantasy, Inc., 510 U.S. 517, 527 (1994).

    39.    Id. at 503.  The Bose rule applies equally to jury trials and bench trials.  466 U.S. at 508 & n.27 (citing New York Times Co. v. Sullivan, 376 U.S. 254 (1964)).  New York Times Co. v. Sullivan specifically held that the Seventh Amendment's ban on "re-examin[ation]" of "fact[s] tried by a jury" didn't preclude independent review by appellate courts in constitutional cases.  376 U.S. at 285 & n.26.

    40.    See Whelan Associates, Inc. v. Jaslow Dental Laboratory, Inc., 797 F.2d 1222, 1233 (3rd Cir. 1986); Hennon v. Kirkland's Inc., 1995 WL 490266 (4th Cir.) (unpublished); Kepner-Tregoe, Inc. v. Leadership Software, Inc., 12 F.3d 529, 534 (5th Cir. 1994); Wildlife Express Corp. v. Carol Wright Sales, Inc., 18 F.3d 502, 506 (7th Cir. 1994); Williams v. Kaag Mfrs, Inc., 338 F.2d 949, 951 (9th Cir. 1964).

    41.    See Concord Fabrics, Inc. v. Marcus Brothers Textile Corp., 409 F.2d 1315 (2nd Cir. 1969); Sherry Mfg. Co. v. Towel King of Florida, Inc., 753 F.2d 1565, 1569 n.6 (11th Cir. 1985).  But see Original Appalachian Artworks, Inc. v. Toy Loft, Inc., 684 F.2d 821, 824 n.4 (11th Cir. 1982) (review for clear error).  See also Mitch Holdings, Inc. v. Arce Engineering Co., 89 F.3d 1548 (11th Cir. 1996) (holding that the idea/expression determination is a mixed question of law and fact; the Eleventh Circuit generally reviews such questions de novo, see International Ins. Co. v. Johns, 874 F.2d 1447, 1453 (11th Cir. 1989), but Mitch Holdings did not clearly indicate the standard of review that it was applying).

    42.    See, e.g., CMM Cable Rep, Inc. v. Ocean Coast Properties, Inc., 97 F.3d 1504, 1525 (1st Cir. 1996); Ford Motor Co. v. Summit Motor Products, Inc., 930 F.2d 277, 290 (3rd Cir. 1991); Gaste v. Kaiserman, 863 F.2d 1061, 1068-69 (2nd Cir. 1988); Transgo, Inc. v. Ajac Transmission Parts Corp., 768 F.2d 1001, 1013-14 (9th Cir. 1985).

              As we mentioned supra in note 7, we discuss here substantial similarity, rather than probative similarity.

    43.    Indeed, this is probably right for probative similarity, though not for substantial similarity.  The failure of many courts to clearly distinguish these two kinds of similarity may help explain why they use clear error review for both.

    44.    See American Geophysical Union v. Texaco Inc., 37 F.3d 881, 886 (2d Cir. 1994); North River Ins. Co. v. Cigna Reinsurance Co., 52 F.3d 1194, 1218 (3d Cir. 1995); Waters v. Commissioner, 48 F.3d 838, 842 (4th Cir. 1995); Davis v. Odeco, Inc., 18 F.3d 1237, 1245 n.30 (5th Cir. 1994); United States v. Clark, 982 F.2d 965, 968 (6th Cir. 1993); Cooper Tire & Rubber Co. v. St. Paul Fire & Marine Ins. Co., 48 F.3d 365, 369 (8th Cir. 1995); Jordan v. Clark, 847 F.2d 1368, 1375 & n.7 (9th Cir. 1988) (de novo review whenever question "requires [the appellate court] to consider legal concepts in the mix of fact and law and to exercise judgment about the values that animate legal principles"); Ershick v. United Missouri Bank, 948 F.2d 660, 666 (10th Cir. 1991); International Ins. Co. v. Johns, 874 F.2d 1447, 1453 (11th Cir. 1989); Carter v. Bennett, 840 F.2d 63, 65 (D.C. Cir. 1988).  But see Williams v. Poulos, 11 F.3d 271, 278 & n.11 (1st Cir. 1993) (mixed questions of law and fact reviewed with varying degrees of deference, depending on how "fact dominated" the question is); Mars Steel Corp. v. Continental Bank N.A., 880 F.2d 928, 933 (7th Cir. 1989) (en banc) ("fact-bound" mixed questions of law and fact reviewed for clear error).  The Supreme Court has not resolved this question.  Pullman-Standard Co. v. Swint, 456 U.S. 273, 289 n.19 (1982).

    45.    See, e.g., MacArthur v. University of Tex. Health Ctr., 45 F.3d 890, 896 (5th Cir. 1995); Baskerville v. Culligan Int'l Co., 50 F.3d 428, 431 (7th Cir. 1995); Ingram v. Acands, Inc., 977 F.2d 1332, 1340 (9th Cir. 1992); Quick v. Peoples Bank, 993 F.2d 793, 797 (11th Cir. 1993); Therma-Tru Corp. v. Peachtree Doors Inc., 44 F.3d 988, 991 (Fed. Cir. 1995); Meyers v. Chapman Printing Co., 840 S.W.2d 814, 822-23 (Ky. 1992).

    46.    The Bose rule applies to jury trials as well as bench trials.  See supra note 39.

    47.    Ornelas v. United States, 116 S. Ct. 1657, 1662, 1663 (1996).

    48.    Harte-Hanks Communications, Inc. v. Connaughton, 491 U.S. 657, 688-89 (1989); Bose Corp. v. Consumers Union, 466 U.S. 485, 499-500 (1984).

    49.    Indeed, this is the rationale underlying the Second Circuit's de novo review for substantial similarity findings.  See Concord Fabric, Inc. v. Marcus Brothers Textile Corp., 409 F.2d 1315, 1317 (2nd Cir. 1969) ("As we have before us the same record, and as no part of the decision below turned on credibility, we are in as good a position to determine the question as is the district court.").

    50.    See, e.g., Bose Corp. v. Consumers Union, 466 U.S. 485 (1984); United States v. D'Ambrosio, No. 92-10526, 1993 WL 41054 (9th Cir. Oct. 14, 1993) (unpublished) ("the standard of review controls the outcome of this case"); United States v. Conley, 4 F.3d 1200, 1204 (3d Cir. 1993) ("the standard of review can be outcome determinative"); Payne v. Borg, 982 F.2d 335, 338 (9th Cir. 1993) ("[t]he relevant standards of review are critical to the outcome of this case"); United States v. Vontsteen, 950 F.2d 1086, 1091 (5th Cir. 1992) (en banc) ("the standard chosen often affects the outcome of the case"); Paul R. Michel (Circuit Judge on the Court of Appeals for the Federal Circuit), Advocacy in the Federal Circuit, C961 A.L.I.-A.B.A. 5, *8 (1994) ("One of my main messages to you is that standards of review influence dispositions in the Federal Circuit far more than many advocates realize."); William H. Kenety, Observations on Teaching Appellate Advocacy, 45 J. Legal Ed. 582, 586 (1995) ("[t]he applicable standard of review determines the outcome of many appellate decisions"); Sally Baumler, Appellate Review Under the Bail Reform Act, 1992 U. Ill. L. Rev. 483, 486 ("[b]ecause the standard of review can affect the outcome of a case, one of the first issues in any appeal is the proper standard of appellate review to be applied"); W. Wendell Hall, Standards of Appellate Review in Civil Cases, 21 St. Mary's L.J. 865, 867 (1990) ("[b]ecause the appropriate standard of review will control the outcome of an appeal, appellate practitioners must consider the standard of review with the same thoughtful consideration that they give to the facts and the substantive law"); see also Fed. R. App. P. 28(a)(6) (requiring appellants to brief the standard of review); United States v. McConney, 728 F.2d 1195, 1199-1204 (9th Cir. 1984) (en banc) (discussing standards of review at length); cf. Michael Asimow, The Scope of Judicial Review of Decisions of California Administrative Agencies, 42 UCLA L. Rev. 1157, 1189 n.112 (1995) ("As an extreme example, one practitioner told me that in many years of practice representing professional licensees . . . he had never lost an independent judgment case and never won a substantial evidence case.").

    51.    Trial court cases are never binding precedent; and if the case is tried to a jury, or is tried to a judge but the decision isn't reported, the case won't even be persuasive precedent.  Even if the decision comes in a bench trial and gets published, it can be persuasive precedent only in future bench trials.  In future jury trials, the trial court won't be able to use the prior decision to grant summary judgment -- without independent judgment review, the question of substantial similarity is for the jury.  And the jury will never know about the past trial court decision.

    52.    Layman v. Combs, 994 F.2d 1344, 1355 (9th Cir. 1992) (Kozinski, J., dissenting) (discussing nature of review under a "no rational trier of fact" standard in a similar context).  Of course, this does sometimes happen.  See, e.g., Kouf v. Walt Disney Pictures & Television, 16 F.3d 1042 (9th Cir. 1994) (movie "Honey, I Shrunk the Kids" did not infringe screenplay called "The Formula"); Berkic v. Crichton, 761 F.2d 1289 (9th Cir. 1985) (move "Coma" did not infringe treatment called "Reincarnation, Inc."); Litchfield v. Spielberg, 736 F.2d 1352 (9th Cir. 1984) (movie "E.T." did not infringe musical play "Lokey from Maldemar").  Summary judgment in favor of the plaintiff is rarer, but does happen occasionally.  See, e.g., Rogers v. Koons, 960 F.2d 301 (2nd Cir. 1992) (sculpture "String of Puppies" infringed photograph "Puppies").

    53.    See Bartimo v. Horsemen's Benevolent & Protective Ass'n, 771 F.2d 894, 897 (5th Cir. 1985); Don's Porta Signs, Inc. v. City of Clearwater, 892 F.2d 1051, 1053-54 n.9 (11th Cir. 1987); see also Lindsay v. City of San Antonio, 821 F.2d 1103, 1107-08 (5th Cir. 1987) (applying independent judgment review even though the free speech claimant won below; no discussion of whether the standard should be symmetrical); Hardin v. Santa Fe Reporter, Inc., 745 F.2d 1323 (10th Cir. 1984) (same).

    54.    See Multimedia Publishing Co. v. Greenville-Spartanburg Airport Dist., 991 F.2d 154, 160 (4th Cir. 1993); Planned Parenthood Ass'n v. Chicago Transit Auth., 767 F.2d 1225, 1229 (7th Cir. 1985); Daily Herald Co. v. Munro, 838 F.2d 380, 383 (9th Cir. 1988); Brown v. K.N.D. Corp., 529 A.2d 1292, 1295-96 (Conn. 1987); see also Don's Porta Signs, Inc. v. City of Clearwater, 485 U.S. 981 (1988) (White, J., dissenting from denial of certiorari) (noting split among the lower courts).

    55.    See Multimedia Publishing Co., 991 F.2d at 160; Planned Parenthood Ass'n, 767 F.2d at 1229; Brown, 529 A.2d at 1295 (similar analysis under state analog of Rule 52(a)).

    56.    See Atlantic & Gulf Stevedores, Inc. v. Ellerman Lines, Ltd., 369 U.S. 355, 358-59 (1962).

    57.    See note 44.

    58.    United States v. McConney, 728 F.2d 1195, 1205 (9th Cir. 1984) (en banc).

    59.    U.S. Const. amend. VII.

    60.    See Mark S. Brodin, Accuracy, Efficiency, and Accountability in the Litigation Process -- The Case for the Fact Verdict, 59 U. Cin. L. Rev. 15, 32, 56-57 (1990); Colleen P. Murphy, Integrating the Constitutional Authority of Civil and Criminal Juries, 61 Geo. Wash. L. Rev. 723, 749 (1993).  But see Comment, Special Verdicts: Rule 49 of the Federal Rules of Civil Procedure, 74 Yale L.J. 483, 502-03 (1965).

    61.    Summary judgment:  See, e.g., Andersen v. McCotter, 100 F.3d 723, 725 (10th Cir. 1996); Price v. Viking Penguin, Inc., 881 F.2d 1426, 1434 (8th Cir. 1989); Secrist v. Harkin, 874 F. 2d 1244, 1251 (8th Cir. 1989); Liberty Lobby, Inc. v. Rees, 852 F.2d 595, 598 (D.C. Cir. 1988); Liberty Lobby, Inc. v. Dow Jones & Co., Inc., 838 F.2d 1287, 1293 (D.C. Cir. 1988) ("First amendment concerns also affect a court's posture in reviewing the evidence presented on summary judgment"; applying Bose on appellate review of summary judgment decision); Herbert v. Lando, 781 F.2d 298, 308 (2d Cir. 1986) (applying Bose on appellate review of summary judgment decision); Coughlin v. Westinghouse Broadcasting & Cable Inc., 780 F.2d 340, 352 n.17 (1986) (Becker, J., concurring); Bartimo v. Horsemen's Benevolent & Protective Ass'n, 771 F.2d 894, 894-98 (5th Cir. 1985); Hardin v. Santa Fe Reporter, Inc., 745 F.2d 1323, 1326 (10th Cir. 1984); Foretich v. ABC, 1997 WL 669644 (D.D.C. Oct. 17) (applying Bose at summary judgment); Davidson v. Time Warner, Inc., 1997 WL 405907, *16 (S.D. Tex. Mar. 31); Rice v. Paladin Enterps., Inc., 940 F. Supp. 836, 844 (D. Md. 1996), rev'd on other grounds, 128 F.3d 233 (4th Cir. 1997); Brown & Williamson Tobacco Corp., 644 F. Supp. 1240, 1245 (N.D. Ill. 1986), aff'd in part and rev'd in part on other grounds, 827 F.2d 1119 (7th Cir. 1987).  But see Liberty Lobby, Inc. v. Anderson, 746 F.2d 1563, 1571 (D.C. Cir. 1984) (taking the opposite view), rev'd on other, but related, grounds, 477 U.S. 242 (1986), and since abandoned by the D.C. Circuit, Liberty Lobby, Inc. v. Dow Jones & Co., Inc., 838 F.2d 1287, 1293 (D.C. Cir. 1988); Coughlin v. Westinghouse Broadcasting & Cable, Inc., 603 F. Supp. 377, 388 (E.D. Pa. 1985) (same, citing Liberty Lobby, 746 F.2d 1563), aff'd on other grounds, 780 F.2d 340 (1986).

              Some of these cases apply Bose on appellate review of a trial court's summary judgment decision, but as we discuss below it would make no sense to apply a different standard to the trial court's decision itself.

              J.n.o.v.:  See Crowder v. Housing Authority, 990 F.2d 586, 594 n.15 (11th Cir. 1993) ("plaintiff was entitled -- under ordinary Federal Rules standards (and even more in the light of Bose's admonition to judges about mixed questions of law and fact) -- to a judgment as a matter of law on most of his constitutional claims); Masson v. New Yorker Magazine, 832 F. Supp. 1350, 1355 (N.D. Cal. 1993); Guccione v. Hustler Magazine, Inc., 632 F. Supp. 313, 317 (S.D.N.Y. 1986), rev'd on other grounds, 800 F.2d 298 (2nd Cir. 1986).


              It is, of course, possible that the district court's independent review would reach one result and the court of appeals' independent review would reach another; in retrospect, then, one might say that the district court's independent review was a waste of time.  But unless the substantial similarity of expression test is indeed entirely indeterminate, we'd assume that by and large the district court and the court of appeals would come to the same, and one hopes correct, conclusion.

    63.    Those skeptical about this might check out any copyright casebook or treatise, and see how many of the leading cases discussed there are district court cases.

    64.    Id.

    65.    See, e.g., Waters v. Churchill, 114 S. Ct. 1878, 1889 (1994) (plurality) ("cases cannot be read as foreclosing an argument that they never dealt with"); Miller v. California Pac. Medical Ctr., 991 F.2d 536, 541 (9th Cir. 1993).

    66.    Compare Dallas Cowboys Cheerleaders, Inc. v. Scoreboard Posters, Inc., 600 F.2d 1184, 1187 (5th Cir. 1979) (concluding that prior restraint doctrine doesn't apply to copyright and trademark cases because "The first amendment is not a license to trammel on legally recognized rights in intellectual property") with L.L. Bean, Inc. v. Drake Publishers, Inc., 811 F.2d 26, 28-29 (1st Cir. 1986) (disagreeing with this in the trademark context; "the constitutional issue raised here cannot be dispensed with by simply asserting that Bean's property right need not yield to the exercise of first amendment rights").

    67.    Reputation is generally not a property interest for purposes of the federal Due Process Clause, Paul v. Davis, 424 U.S. 693 (1976), but it may be a property right for other purposes.  E.g., Marrero v. City of Hialeah, 625 F.2d 499, 514 (5th Cir. 1980) (Florida law recognizes business reputation as a property interest); Nossen v. Hoy, 750 F. Supp. 740, 743 (E.D. Va. 1990) ("an individual holds a . . . property interest in his or her reputation" for purposes of Washington and Virginia conversion law).

    68.    See San Francisco Arts & Athletics, Inc. v. United States Olympic Committee, 483 U.S. 522 (1987) (quasi-trademark); Harper & Row (copyright); Zacchini v. Scripps-Howard Broadcasting Co., 433 U.S. 562 (1977) (right of publicity).

    69.    See New York Times Co. v. Sullivan, 376 U.S. 254, 265 (1964) ("Although [a libel suit] is a civil lawsuit between private parties, the [state] courts have applied a state rule of law . . . .  It matters not that that law has been applied in a civil action . . . .").

    70.    And query whether privately enforced laws might actually prove to be more restrictive than government-enforced ones.  See Hughes Aircraft Co. v. United States, 117 S. Ct. 1871 (1997) (suggesting that private, self-interested enforcement of legal rules may be more zealous and more thorough than direct government enforcement).

    71.    See Dr. Seuss Enterprises, L.P. v. Penguin Books USA, Inc., 109 F.3d 1394, 1403 n.11 (9th Cir. 1997), quoting Dallas Cowboys Cheerleaders, Inc. v. Pussycat Cinema, Ltd., 604 F.2d 200, 206 (2d Cir. 1979) (holding that trademark law is content-neutral).

    72.    Turner Broadcasting System, Inc. v. FCC, 114 S. Ct. 2445, 2471 (1994); Association of Community Orgs. for Reform Now v. St. Louis County, 930 F.2d 591, 595-96 (8th Cir. 1991).

    73.    See, e.g., Burson v. Freeman, 112 S. Ct. 1846, 1850 (1992) (plurality) (holding that all content-based restrictions, even viewpoint-neutral ones, are constitutionally suspect); Boos v. Barry, 485 U.S. 312, 321 (1988) (same); Arkansas Writers' Project, Inc. v. Ragland, 481 U.S. 221, 230 (1987) (same); Consolidated Edison Co. v. Public Serv. Comm'n, 447 U.S. 530, 537-38 (1980) (same); Carey v. Brown, 447 U.S. 455, 462 n.6 (1980) (same).

    74.    See Twentieth Century-Fox Film Corp. v. MCA, Inc., 715 F.2d 1327 (9th Cir. 1983).

    75.    See supra note 29.

    76.    In fact, in obscenity and fighting words cases, the risk of error is borne by speech that at least some Justices have claimed is of low constitutional value.  Compare Young v. American Mini Theatres, Inc., 427 U.S. 50 (1976) (plurality) (suggesting that pornography is of low constitutional value) and FCC v. Pacifica Foundation, 438 U.S. 726 (1978) (suggesting that profanity is of low constitutional value) with R.A.V. v. City of St. Paul, 112 S. Ct. 2538 (1992) (stressing that this view has never commanded a majority of the Court).  Entertainment, even nonpolitical entertainment, has always been held to be of high constitutional value.  See Winters v. New York, 333 U.S. 506, 510 (1948).

    77.    See supra note 36.

    78.    Cf. Time, Inc. v. Firestone, 424 U.S. 448 (1976) (involving a libel action based on a story about the divorce of a wealthy socialite); see also Time, Inc. v. Hill, 385 U.S. 374 (1967) (involving a false light privacy action based on a sensational crime story that's more akin to a modern docudrama than to political speech); Cantrell v. Forest City Pub. Co., 419 U.S. 245 (1974) (same).

    79.    Cf. Frederick Schauer, Uncoupling Free Speech, 92 Colum. L. Rev. 1321 (1992) (suggesting that the chilling effect of libel law on media businesses is less than one might think).

    80.    Harper & Row, 471 U.S. at 558.  Cf. Patton v. United States, 281 U.S. 276, 298 (1930) ("The first ten amendments and the original Constitution were substantially contemporaneous and should be construed in pari materia.").

    81.    See Harper & Row, 471 U.S. at 558.

    82.    Cf. Fogerty v. Fantasy, Inc., 510 U.S. 517, 527 (1994) ("a successful defense of a copyright infringement action may further the policies of the Copyright Act every bit as much as a successful prosecution of an infringement claim by the holder of a copyright").

    83.    Gertz v. Robert Welch, Inc., 418 U.S. 323, 399-400 (1974) (White, J., dissenting) (also endorsing the view that "fascists' effective use of defamatory attacks on their opponents" suggests that "the law of libel . . . [may be] "important for modern democratic survival").

    84.    See, e.g., Andrea Dworkin & Catharine A. MacKinnon, Pornography and Civil Rights: A New Day for Women's Equality 47-48 (1988); Paul E. McGreal, Constitutional Illiteracy, 30 Ind. L. Rev. 693, 697 n.30 (1997) (discussing this argument).

    85.    Kunz v. New York, 340 U.S. 290, 302 (11951) (Jackson, J., dissenting).

    86.    Cf. Cable/Home Communication Corp. v. Network Productions, Inc., 902 F.2d 829, 849 (11th Cir. 1996) ("Since the Copyright Act is the congressional implementation of a constitutional directive to encourage inventors by protecting their exclusive rights in their discoveries, copyright interests also must be guarded under the Constitution, and injunctive relief is a common judicial response to infringement of a valid copyright"); 1 Rodney A. Smolla, Smolla and Nimmer on Freedom of Speech § 15:58, at 15-87 to -88 ("the fact that copyright protection is itself a value of constitutional dimension, vindicating the directive of the Constitution's Copyright Clause," injunctions should be permitted).

    87.    See, e.g., Lamont v. Postmaster General, 381 U.S. 301, 307 (1965) (striking down a restriction on mailing of Communist advocacy); see also, e.g., United States v. Grace, 461 U.S. 171, 179-80 (1983) (striking down on First Amendment grounds a law enacted pursuant to the Federal District Clause power); Alex Kozinski & Eugene Volokh, A Penumbra Too Far, 106 Harv. L. Rev. 1639, 1649-50 (1993).

    88.    See, e.g., Time Warner Entertainment Corp. v. Does, 876 F. Supp. 407 (S.D.N.Y. 1994) (holding that ex parte seizure of supposedly infringing materials, though authorized by the Copyright Act, was impermissible under the Fourth Amendment).

    89.    See Patton v. United States, 281 U.S. 276, 298 (1930) ("The first ten amendments and the original Constitution were substantially contemporaneous and should be construed in pari materia.").

    90.    Harper & Row, 471 U.S. at 558.

    91.    But see Religious Technology Center v. Netcom On-Line Comm. Servs., Inc., 923 F. Supp. 1231, 1258 (N.D. Cal. 1995) (concluding that Harper & Row implicitly validated all the provisions of the Copyright Act, including those providing for preliminary injunction).

    92.    See, e.g., Milkovich v. Lorain Journal Co., 497 U.S. 1, 22 (1990).

    93.    See, e.g., New York v. Ferber, 458 U.S. 747 (1982).

    94.    Peter Pan Fabrics, Inc. v. Martin Weiner Corp., 274 F.2d 487, 489 (2d Cir. 1960).

    95.    630 F.2d 905 (2nd Cir. 1980).

    96.    Id. at 917, quoting Couleur International Ltd. v. Opulent Fabrics, Inc., 330 F.Supp. 152, 153 (S.D.N.Y. 1971).

    97.    Id.

    98.    Id. at 916.

    99.    Cf. Past Pluto Productions Corp. v. Dana, 627 F. Supp. 1435, 1441 (S.D.N.Y. 1986) (similarly relying on several cases in determining copyrightability, as opposed to substantial similarity).

    100.    712 F. Supp. 353 (S.D.N.Y. 1989).

    101.    865 F. Supp. 1047 (S.D.N.Y. 1994).

    102.    945 F.2d 509 (2nd Cir. 1991).

    103.    736 F.2d 859 (2nd Cir. 1984).

    104.    937 F.2d 700 (2nd Cir. 1991).

    105.    499 U.S. 340 (1991).

    106.    899 F.2d 1537 (7th Cir. 1990).

    107.    Id. at 1540.

    108.    Id. at 1541.

    109.    618 F.2d 972 (2nd Cir. 1980).

    110.    181 F.2d 664 (7th Cir. 1950).

    111.    See, e.g., Herbert Rosenthal Jewelry Corp. v. Honora Jewelry Co., 509 F.2d 64 (2nd Cir. 1974) (drawing on Herbert Rosenthal Jewelry Corp. v. Kalpakian, 446 F.2d 738 (9th Cir. 1971)); Uneeda Doll Co. v. Regent Baby Products Corp., 355 F. Supp. 438 (E.D.N.Y. 1972) (drawing on Ideal Toy Corp. v. Fab-Lu Ltd., 360 F.2d 1021 (2nd Cir. 1966)); Knitwaves, Inc. v. Lollytogs Ltd., 71 F.3d 996 (2nd Cir. 1995) (extensive comparison and contrast with Folio Impressions, Inc. v. Byer California, 937 F.2d 759 (2nd Cir. 1991)).

    112.    Bose, 466 U.S. at 502-04.

    113.    Fogerty v. Fantasy, Inc., 510 U.S. 517, 527 (1994).

    114.    The practitioners to whom we talked are:  Blaine Greenberg of Troop, Meisinger, Steuber and Pasich in Los Angeles (April 10, 1997 Phone Interview); Tom Hemnes of Foley, Hoag, and Eliot in Boston (April 21, 1997 Phone Interview); David Nimmer of counsel to Irell & Manella in Los Angeles (March 5, 1996 Phone Interview); Peter Nolan of the Walt Disney Company in Los Angeles (March 11, 1997 Phone Interview); Bob Osterberg of Abelman, Frayne and Schwab in New York (March 25, 1997 Phone Interview); and Herb Schwartz of Fish and Neave in New York (April 16, 1997 Phone Interview).

    115.    For the full questionnaire, see the Appendix.

    116.    See Appendix, question 1.

    117.    Nimmer, Hemnes, and Nolan Interviews, supra note 114.

    118.    Osterberg and Schwartz Interviews, supra note 114.

    119.    Greenberg Interview, supra note 114.

    120.    See Appendix, question 2.

    121.    Osterberg Interview, supra note 114.

    122.    Schwartz Interview, supra note 114.

    123.    Greenberg Interview, supra note 114.

    124.    See Appendix, question 4.  Greenberg had the interesting suggestion of a database that would list not only appellate decisions by area, but would also lay side-by-side comparisons of the material being litigated.  Greenberg Interview, supra note 114.

    125.    Schwartz Interview, supra note 114, alluding to Justice Stewart's famous statement that "I shall not today attempt further to define [hard-core pornography]; and perhaps I could never succeed in intelligibly doing so.  But I know it when I see it . . . ."  Jacobellis v. Ohio, 378 U.S. 184, 197 (1964) (Stewart, J., concurring).

    126.    See Appendix, question 5.

    127.    Hemnes Interview, supra note 75.

    128.    Nolan Interview, supra note 114.

    129.    Greenberg Interview, supra note 114.

    130.    Osterberg Interview, supra note 114.

    131.    Schwartz Interview, supra note 114.

    132.    Nimmer Interview, supra note 114.

    133.    See Appendix, question 6.

    134.    Greenberg Interview, supra note 114.

    135.    Osterberg Interview, supra note 114.

    136.    See Appendix, question 7.

    137.    Nimmer Interview, supra note 114.

    138.    Greenberg Interview, supra note 114.

    139.    Schwartz Interview, supra note 114.

    140.    See Appendix, question 8.

    141.    Hemnes Interview, supra note 75.

    142.    Osterberg Interview, supra note 114.

    143.    See Appendix, question 11.

    144.    See Appendix, question 10.

    145.    Hemnes Interview, supra note 75.

    146.    See Appendix, question 9.

    147.    Greenberg Interview, supra note 114.

    148.    Nimmer Interview, supra note 114.

    149.    Nolan Interview, supra note 114.

    150.    116 S. Ct. 1657 (1996).

    151.    Id. at 1661 (citations omitted).

    152.    Id. at 1662-63 (citations omitted).

    153.    116 S. Ct. 457 (1995).

    154.    Id. at 465 n.11.

    155.    Id. at 466 n.13.

    156.    Id. at 466 & n.13.

    157.    Id. at 467.

    158.    See Cooter & Gell v. Hartmarx Corp., 496 U.S. 384 (1990), which dealt with the question of whether a lawyer engaged in reasonable investigation for purposes of Rule 11 of the Federal Rules of Civil Procedure; and Pierce v. Underwood, 487 U.S. 522 (1988), which dealt with whether the government's position in litigation was substantially justified for purposes of determining Equal Access to Justice Act fee awards.  In both cases, the Court held that deferential review was required, because the fact patterns involve "multifarious, fleeting, special, narrow facts that utterly resist generalization" that independent appellate review could not "clarify the underlying principles of law."  Cooter & Gell, 496 U.S. at 404-05; Pierce, 487 U.S. at 560-62.  Unfortunately, the Court hasn't explained exactly why it was treating Ornelas and Thompson differently from Cooter & Gell and Pierce; the best explanation, though, seems to be that the latter cases involved matters that were peripheral to the merits, and that were not constitutionally significant.

    159.    See Miller v. Fenton, 474 U.S. 104, 114, 116-17 (1985) (citing cases that mandate deferential review in certain habeas corpus contexts).

    160.    Of course, decisions about whether the defendant used plaintiff's work or whether defendant's work is an entirely independent creation do often turn on credibility judgments; still, this is an analytically separate inquiry from the decision whether plaintiff's expression is substantially similar to defendant's expression.

    161.    Ornelas, 116 S. Ct. at 1661.

    162.    See, e.g., Hemnes comment, note 145 and accompanying text; and compare Second Circuit factual compilation cases such as Eckes v. Card Prices Update, 736 F.2d 859 (2nd Cir. 1984), and Key Publications, Inc. v. Chinatown Today Publishing Enterprises, 945 F.2d 509 (2nd Cir. 1991), with Ninth Circuit cases such as Landsberg v. Scrabble Crossword Game Players, Inc., 736 F.2d 485 (9th Cir. 1984), Cooling Systems & Flexibles, Inc., 777 F.2d 485 (9th Cir. 1985), and Transgo, Inc. v. Ajac Transmission Parts Corp., 768 F.2d 1001 (9th Cir. 1985); cf. also the contrast between Durham Industries, Inc. v. Tomy Corp., 630 F.2d 905 (2nd Cir. 1980), and Williams v. Kaag Mfrs., Inc., 338 F.2d 949 (9th Cir. 1964).

    163.    See Harper & Row Publishers, Inc. v. Nation Enters., 471 U.S. 439 (1985); Sega Enters. LTD v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992); Acuff-Rose Music, Inc. v. Campbell, 972 F.2d 1429, 1434 (6th Cir. 1992), rev'd on other grounds, 510 U.S. 569 (1994); New Era Publications v. Carol Pub. Group, 904 F.2d 152, 155 (2nd Cir. 1990); Pacific & Southern Co. v. Duncan, 744 F.2d 1490, 1495 n.8 (11th Cir. 1984); Triangle Publications, Inc. v. Knight-Ridder Newspapers, Inc., 626 F.2d 1171, 1175 (5th Cir. 1980).

    164.    Maxtone-Graham v. Burtchaell, 803 F.2d 1253, 1254 (2d Cir. 1986); see also Triangle Publications, Inc. v. Knight-Ridder Newspapers, Inc., 626 F.2d 1171, 1174 (5th Cir. 1980).

    165.    Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 596 (1994) (Kennedy, J., concurring).

    166.    464 U.S. 417 (1984).

    167.    510 U.S. 569, 596 (1994).

    168.    Bose, 466 U.S. at 502.

    169.    Martin E. Lee, The Price We Pay: The Case Against Racist Speech, Hate Propaganda and Pornography, Nat'l Catholic Rep., Oct. 4, 1996, at 17 (book review) ("Noting routine exceptions to free speech absolutism (copyright, trademark and such) that hew to business interests, the essays cite studies that document the heavy toll inflicted by the multibillion dollar porn industry, as it profits from a kind of hate speech that degrades women and children. . . .  This book provides a sober rejoinder to cliche-ridden thinking by highlighting the profound power imbalance and social inequities that dim the luster of the First Amendment."); Richard Delgado & Jean Stefancic, Ten Arguments Against Hate-Speech Regulation: How Valid?, 23 N. Ky. L. Rev. 475, 484 (1996) ("Powerful actors like government agencies, the writers' lobby, industries, and so on have always been successful at coining free speech `exceptions' to suit their interest -- copyright, false advertising, words of threat, defamation, libel, plagiarism, words of monopoly, and many others.  But the strength of the interest behind these exceptions seems no less than that of a black undergraduate subjected to vicious abuse while walking late at night on campus."); Richard Delgado & David H. Yun, Pressure Valves and Bloodied Chickens: An Analysis of Paternalistic Objections to Hate Speech Regulation, 82 Calif. L. Rev. 871, 892 (1994) ("Perhaps . . . in twenty or fifty years we will look upon hate speech rules with the same equanimity with which we now view defamation, forgery, obscenity, copyright, and dozens of other exceptions to the free speech principle, and wonder why in the late twentieth century we resisted them so strongly.").

    170.    See United States v. McConney, 728 F.2d 1195, 1201 n.7 (9th Cir. 1984) (en banc) ("It can hardly be disputed that application of a non-deferential standard of review requires a greater investment of appellate resources than does application of the clearly erroneous standard.  Appellate courts could do their work more quickly if they applied the clearly erroneous standard in most circumstances, because the courts then need only determine if the lower court's decision is a reasonable one, not substitute their own judgment for that of the trial judge.").

    171.    We suspect that plaintiffs and defendants will differ more in their estimates of success under independent review than in their estimates of success under deferential review.  If this is so, then settlement will be less likely.  See United States v. McKinney, 919 F.2d 405, 422 (7th Cir. 1990) (Posner, J., concurring); Fed. R. Civ. Proc. 52(a), advisory committee note (independent appellate review "tend[s] to . . . multiply appeals by encouraging appellate retrial of some factual issues").

    172.    Speiser v. Randall, 357 U.S. 513, 526 (1957).

    173.    Frequent communications from Vladimir Volokh to Eugene Volokh.

    174.    Cf., e.g., N.Y. Const. art. VII, § 8 (1821) ("In all prosecutions or indictments for libels, the truth may be given in evidence to the jury; and if it shall appear to the jury that the matter charged as libellous is true, and was published with good motives and for justifiable ends, the party shall be acquitted").

    175.    New York Times Co. v. Sullivan, 376 U.S. 254 (1964).

    176.    Cf, e.g., Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 60 (1971).

    177.    Gertz v. Robert Welch, Inc., 418 U.S. 323, 390 (1974) (White, J., dissenting).

    178.    Id.

    179.    Id. at 395 & n.31.

    180.    Bose, 466 U.S. at 519 (Rehnquist, J., dissenting).

    181.    Bates v. State Bar, 433 U.S. 350, 380 (1977) ("First Amendment interests are fragile interests, and a person who contemplates protected activity might be discouraged by the in terrorem effect of the statute."); NAACP v. Button, 371 U.S. 415, 433 (1963) ("These freedoms are delicate and vulnerable . . . .  The threat of sanctions may deter their exercise almost as potently as the actual application of sanctions.").

    182.    Southeastern Promotions, Ltd. v. Conrad, 420 U.S. 546, 560 (1975); Blount v. Rizzi, 400 U.S. 410, 416 (1971); Bantam Books, Inc. v. Sullivan, 372 U.S. 58, 66 (1963); see also FW/PBS, Inc. v. City of Dallas, 493 U.S. 215, 229 (1990).

              Another alternative, proposed by Professor Monaghan in his Constitutional Fact Review, supra note 20, is for courts to use independent review until they believe they've refined the test as much as they can, and then shift to reviewing only for clear error.  We're not sure this approach will ultimately work, because it's hard for courts to tell when enough is enough.  The standard of review will be essentially up in the air for a long time, with the appellant in each case insisting that there's more refinement possible and thus asking for de novo review, and with the appellee arguing that it's now time to switch to clear error review.  Still, we agree with Professor Monaghan that Bose might bear some reexamination.

    183.    Perhaps even the manufacturers are entitled to clearer guidelines than the law currently provides; nonetheless, while a considerable amount of uncertainty might have to be tolerated for the manufacturers, we believe that less should be tolerable for speakers.

    184.    Speiser v. Randall, 357 U.S. 513, 526 (1957).

    185.    See Vance v. Universal Amusement Co., 445 U.S. 308, 312 (1980).

    186.    Cf. Mark Lemley & Eugene Volokh, Freedom of Speech and Preliminary Injunctions in Intellectual Property Cases (in progress).

    187.    Gertz v. Robert Welch, Inc., 418 U.S. 323 (1974).

    188.    Smith v. California, 361 U.S. 147 (1959) (criminal cases); Manual Enterprises, Inc. v. Day, 370 U.S. 478 (1962) (civil cases).

    189.    New York v. Ferber, 458 U.S. 747 (1982).  Of course, some statutes allow any mistake as a defense, even an unreasonable one, see, e.g., United States v. X-Citement Video, Inc., 115 S. Ct. 464 (1994); it's unclear, though, whether any mens rea beyond negligence is constitutionally required.

    190.    Cf. De Acosta v. Brown, 146 F.2d 408, 412 (2d Cir. 1944) (Hand, J., dissenting) (arguing that holding magazine publisher strictly liable for infringement by contributing author "is likely to prove an appreciable and very undesirable burden upon the freedom of the press"); Eugene Volokh, Cheap Speech and What It Will Do, 104 Yale L.J. 1805, 1844 n.130 (1995) (briefly touching on this question); Edward M. Di Cato, Operator Liability Associated with Maintaining a Computer Bulletin Board, 4 Software L.J. 147, 155-56 (1990) (discussing this question).

    191.    Gertz.

    192.    17 U.S.C. § 504.  Statutory damages are conventionally seen more as presumed damages than punitive, but some cases suggest that they also have a punitive component.  See, e.g., Cass County Music Co. v. C.H.L.R., Inc., 88 F.3d 635 (8th Cir. 1996) ("it is plain that another role has emerged for statutory damages in copyright infringement cases: that of a punitive sanction on infringers," akin to "the award of punitive damages"); Evans Newton Inc. v. Chicago Systems Software, 793 F.2d 889 (7th Cir. 1986) ("Under copyright law, punitive damages could come from an award of statutory damages for willful infringement."); Video Cafe, Inc. v. De Tal, 961 F. Supp. 23 (D.P.R. 1997) ("The Court would also note that statutory damages awards under § 504(c) serve both compensatory and punitive purposes.").

    193.    Philadelphia Newspapers v. Hepps, 475 U.S. 767 (1986).

    194.    Harper & Row, 471 U.S. at 560.

    195.    Cf. Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 590 (1994) (fair use is an affirmative defense, and the burden of proving it is on the defendant).

    196.    Gertz v. Robert Welch, Inc., 418 U.S. 323, 342 (1974).

    197.    See People v. Mitchell Brothers' Santa Ana Theater, 128 Cal.App.3d 937, 940, 180 Cal.Rptr. 728, 730, 127 Cal. Rptr. 317, 545 P.2d 229 (1982).

    198.    See Yen, supra note 27, at 434-35 (mentioning this in passing).

    199.    Stare decisis doesn't prevent courts from adopting this approach, even if they have in the past reviewed substantial similarity findings only for clear error.  See supra Part I.E.